I
Writers
In some months, SCRIVENER will be here, in the Windows world. Scrivener is a myth for Windows users, and many Apple users write novels, screenplays, stage plays with it... some Windows users have long ago bought a Mac just for being able to use that splendid software.
Why? It's functional for writers, and it's very beautiful indeed, and people LIKE to use it.
Have any (fictional) writer. He writes in "chronological" order, using a tree, an outline, and even when he breaks up chronological order of his story, he will maintain what I would like to put as the "chronological-reading-order", or in film, the "screen order", tce same on stage, the chronological succession of stage action, be there multiple flashbacks or just "chronological-order" story-telling.
In all this, he will employ his personnel, let's say th = Thomas, ma = Mary, ch = Christopher and many others. In Scrivener, in MI also, he will tag his items. The, he will have a look upon those items (=scenes, parts of scenes, alternatives of scenes, being it a screenplay or a novel, no importance) in which th, ma, ch will interact (or, perhaps, become a subject of conversation, or whatsoever, in this case he will not do a tag "th" but "th(con)" or any other).
Let's imagine his work, let's imagine it's a big screenplay, a long novel, these tags "th" will occur a hundred times or more if th is one of the leading personnel of the novel.
He will then try to fraction these th tags, as I said before, th(con) for "here people speak about Thomas", and many more: th(i) for "innver voice", th(d) for "makes a Decision" (= upon which the next 300 pages will depend), or even "th ma" in order to describe a scene in which he's not alone, but in conversation with ma.
So, we have to consider a lot of scenes in which personnel tags are combined, and perhaps, in which those combinations (of several personnel, or of something different) will constitute something new.
Let's consider other tags, our writer will want to tag subjects, themes: ha for happiness, de for despair, es for estrangement and ra for rapprochement, I am speaking here in psychological categories of course, categories affecting the interaction of the personnel.
Of course, there's also the possibility to do rather simple tags, for towns in which our heroes move, de for Detroit, lo for London, whatever.
Our writer will consider this display of tags, 100 for th = Thomas. Then, he wants to know, when th is alone with himself, is he perhaps alone with himself too many chapters in a row, what kind of interaction, with whom, could I introduce here and there, in order to capture interest of my reader.
He will, perhaps, also introduce tags for pace: let's say, p1, p2, p3, where p1 is very detailed writing, slow pacing, p2 is "normal" writing, p3 would be narrating to us 3 months in 3 pages - of course you could do this by an attribute "P" with values from 1 to 3 or 1 to 5, whatever.
I have three points to make here:
a ) It's clear from what I've said above that our writer needs immediate access to tags, and immediate access to tag combinations. Let's analyze this further down.
b ) Our writer needs this access in "chronological order", not in alphabetical one. He will need to analyze the development of his "story", any alphabetical shuffling around of all those (perhaps 100 or more) scenes will not do him any good. If he wants to have this, he can have it, by those filter commands, but 95 times out of 100, he will need to have this filtering by tree order.
c ) He needs tags combinations very often, precisely because if Thomas is in 100 of 120 scenes, looking onto 100 scenes with "th" will not be very helpful most of the time. He COULD do this by the above-mentioned "th(i)", and so on, BUT he then will see that, when he wants to see the scenes with Thomas, he must put SEVERAL tags just for this, "th", "th(i)", "th(x)" and whatever yet - awful, not practical.
So our writers needs to have combinations, looking for "th" AND for "i", here - AND EASY ACCESS to such combinations! -, or, not simple to program, he needs some sort of hierarchical tags, in which looking for "th" will look for ALL "i" combinations, th(i), th(whatever) - WITH the possibility maintained to just looking for th, without any combinations, and finally, it would be necessary to implement the possibility to look for each combination seperately.
What I am trying to explain here is not easy, since I want to say, the MI tagging function would be splendid if it was possible to ADHERE some tags to some other tags. Let me explain, it's an important point.
When you have th, ma and ch, 3 roles, and some themes, let's say lo for loneliness, an AFFILIATION of a given theme to some role - and the possibility to do that affiliation - would make the difference.
Cou could look after th and lo, but this would also give you the scenes where it's ch who's the lonely one, and in the filter list, this Christopher scene would distract, so our writer would like to have only those scenes displayed in his filter view where let's say, there's Mary, Thomas, and one of THEM is lonely, but not third persons.
At this time, you could "program" a filter with "and" and "not" that could display that; it would not be very easy (and error-prone); having the possibility to just have some sort of "th ma lo" tags - but without any lo-tags NOT concerning th and ma - would do it.
I understand this isn't easy, and should be done in a further step.
WHAT is to be done in a first step, that's easy access to (normal) tag combinations, not only "th" OR "ma", but both (or 3, or 4) - and, in any case, in tree order.
In Hollywood, they do 300 (?) films a year, including all independent productions? In the US alone, there must be 3 million would-be screenwriters that are trying with one software or another. THE MARKET IS THERE.
Of course, throwing them MI just at it is, will not make them see the possibilities they would have with MI. Marketing of MI should include some dummy novel, some dummy screenplay, where the columns are displayed when you load the program, and where dozens of tags (and other attributes) are displayed, and where, instead of real fictional text, you'd explain how the tags and other attributes are employed, how they can be helpful in writing - and of course, we would need to have this in some major languages, English, French, German, Spanish perhaps... (I could help in French and German, even a little bit in English, for free, as soon as MI can be considered having a real tough GUI, it's all smoothness of working, I am not asking here and now ANY new feature, just better implementation of the current ones.)
Then, it would be possible to present MI to many writers' associations, many writers' portals in the web, since it would be a real high-brow contender in the Windows world, and WHEN it is, MI would naturally INTEREST all those people managing such portals, magazines... and computer magazines also.
You must see that those computer magazine journalists do NOT have the time to try out a program when you send it to them. They have a look, do NOT understand the various possibilities of it, and lose interest. It's quite another scenario when they open the program, and, instead of just seeing a blank screen, in 3 or more tabs, 3 more topics will open, each with its columns opened, and let's say, one displaying a sketch of a novel, one a sketch of a psychoanalyse, and a third one some hints how to better organize your journalist's work!
THEN those people get interested, get stuck an hour or two with the program... and will be willing to do a little paper in their magazine about it! It's all in the marketing, it's all FREE, but it's a little bit marketing work first, and above all, the GUI has to be perfect to begin with. Here, I am trying to sketch exactly this foundation of MI becoming something really splendid. MI has to CHARM people into their use of it - that's what programs like Scrivener do... and technically, they appear inferior to me...
II
Psychologists
There are some 2,000 psychologists, psychiatrists, psychotherpists and psychoanalysts in Paris alone; troughout the world, there are some millions, I think.
All those people try to organize their work some way or another; most are just working with Microsoft Word, well, let's face it, not all of them do analytical work all of the time, but almost all of them do analytical work at least some time in their work with their "clients" - and then, they are quite lost.
Analytical work is establishing the life of the client. The work isn't done in chronological order, but the tree to be constructed should be in chronological order, and of course, tagging does the work in trying to see themes, recurrent errors, recurrent self-damaging, recurrent behaviors of all sorts - for just one client, estimate some 100 tags just for these, and not including any personnel tagging (which will come on top).
Now, the therapist will constantly look into his tagging system in working with this client, and he will do it so extensively that fluidity in this tagging work will decide if he considers MI or isn't interested to employ it in his work.
As for our fictional writer, he will need to have these tags in combinations, AND in chronological order, i.e. in tree order. (And my remarks above for sophisticated tag combinations apply here, but also, my remark that normal tag combinations would be perfect for the time being, if they can be easily accessed. That very special tag combination thing is NOT necessary to realize before my marketing efforts for MI, promised, let's just have normal tag combination in an easy way, and we'll be happy for the time being!)
III
On TOP of our writer's needs, our psychotherapist / -analyst will need some automatically-driven special "tagging" / stamping system by which NOT just entire items will by tagged - that should be done automatically indeed, on top of that, but first of all, there should be a stamp with the number sign and an automatically-given number, AND that number should automatically be entered as a tag, or in a special attribute field.
Of course, you understand what I am speaking of: the session number! Whenever our psychologist does write down anything, an internal MI routine (WHEN that option is put "on" of course) will look into the tags of that item, or into a special attribute field for that item, and will check if the current session number is there already; if not, it will put it there without asking (and "behind the scenes", without opening any dialog for that: without distraction the psychologist in his work.
Before the session, the psychologist would be informed that MI has opened a new session, and ideally, it would even distinguish real sessions, number sign plus a number, and in-between-the-sessions-reviews (!), where the number sign would be with the last given session number (= and not a new one), but with a special sign after the number, indicating that the psychologist has reviewed this file after that given session, but not within the session.
Problem here, the tag / entry should be numerical altogether, so as to allow for numerical filtering, "from 30th to 50th sessions" or whatever.
This way, in any given session, the psychologist would take notes in a dozen items or so, and then, in each item, above his current writing / paragraph, there would be that session stamp, and in every item with this session stamp, there would be, in addition, that session stamp in the special tagging field. So, if a subject is reviewed again and again, that special field, over time, would be filled up with a lot of session stamps, which is why you cannot use just a numerical field for it (= which would only contain ONE number, instead of a dozen or more).
The point here is, giving to the psychologist the ability to review given sessions, and there should be the possibility to trigger a special session review mode, in which, whenever he clicks on an item belonging to this given session or group of sessions, not only the item would be displayed, but the screen would scroll to that session stamp (or the first relevant session stamp) WITHIN the item.
( In a second step, it should be possible to program a function which REUNITES those stamped passages, from all the concerned items, into a new item; for example, in order to know to which passage the session stamp applies, those psychologists would be invited to just work on paragraphs, AND on several paragraphs NOT seperated by a blank line, when doing their notes; accordingly, MI then would consider all paragraphs NOT seperated by a blank line belonging to that session stamp, and would copy them into that number x session item - of course, in doing this aggregating, MI would do another stamp into each aggregated passage, "stamping" it (=doing an additional line above each passage giving the item's title from which it comes from, and doing a blank line after each passage) - all this, again, will be another step in the future, I think; at this time, for psychologists' professional use, MI just needs the stamping feature, not the aggregation feature.)
It should be noted that the session stamp number needs to be realized by a global variable - trying to realize it by the Windows clipboard would interfere with free use of ^c / ^x / ^v throughout the psychologist's work - but this goes without saying.
III
Now for something tangible since we want to enhance MI, not in the technical way - all technical is there, I said that - but GUI-wise.
a ) I put ShowTags on a function key. There, I need the mouse to click upon ONE tag, and the filter view is shown.
b ) At this time, that's in alphabetical order. As I explained, "we all" (= I, writers, psychologists - AND lawyers also!) need it in tree order... BUT not ALL people do want what we want, so there has to be a choice. Why not doing a right-click mouse menu showing the whole list in order to option for what you want at a given time (= just like the options menu in the (not so easy-to-use, but we'll enhance that another day) filter dialog) - thus, you option for alphabetical, and all that filter display will be alphabetical, up to when you put another option.
c ) I put ShowFilter on another function key, but I see that most of the time, I am going to a), click, and c) then is displayed automatically, so this c) is good to have, but not so helpful. I also put ShowTree on another function key, and indeed, that ShowTree, I use a lot, but again, when I go by ShowTags, click there, ShowFilter is displayed, then I doubleclick there, and the tree is automatically displayed. Oh yes, I forget, to get back to the filter view, I really need that ShowFilter function key!
d ) So I have the tree view, or the filter view, in that pane to the left of my screen: perfect. Perhaps be so good and give us, additionally, a TOGGLE between tree view and filter view.
e ) And since we are at it, give us another TOGGLE between that left pane (whatever its current content) and the editor pane. Since let's face it, the problem with those two commands GoForward/Backwards1Pane is that they are no toggle, and if you do use the GoForward1Pane command to toggle, is does not work but if the command pane (=searches, attributes) is NOT open.
f ) Since we are at it, there is not only necessary that (available) command ShowCommandPane (=a toggle), but we also (=also, not instead) need a command GoToCommandPane (=and open it if it's not yet open of course).
g ) And since we're at it, at this time GoForward1Pane, when in the editor pane, goes not to the tree, but - predictably so - to the command pane, BUT then, it's NOT possible to do anything there, except when you use the mouse, that is to say, "going" to the command pane, with any command (=even by going to it with the search command it's the same problem), gives the focus just to the border of the command pane, but then it should be possible, by tab for example, or even better by DownArrow, to access there the first search result, and then, by doing an Enter, to display that item, and in a general way, it should be possible to go up and down within the search results there by DownArrow and UpArrow, triggering the item display by doing Enter.
h ) As I said before, showing the filter just by clicking on ONE Tag in a ) is fine, but insufficient; easy combinations must be possible. How? By mouse clicks? By Shift- and Control- mouse clicks? Bad! Just do, in addition, a simple command ShowTags which can be put on any function or other key, and which applies automatically to the current item, like the available command ShowTags which displays a list of all available tags. Just that this command would open a tiny dialog where you put the wanted tag(s) into a one-line field, then press Enter, so it would be "F12 ab cd ef Enter" in order to get displayed the ShowFilter with the 3 tags "ab", "cd" and "ef" - I explained elsewhere why tags as short as possible are highly preferable for most of possible uses.
This would be a very important command since it would allow us do have tag combinations displayed without going thru the complicated filter dialog, in fact it would be a very handy additional BASIC version of that filter dialog. The same option as in a ) would apply, i.e. if in the options, you had chosen "alphabetical list", this command would display the concerned items in alphabetical order, and so on. Perhaps you could add the same option field into this dialogue, as is in the filter dialog, and as would be in the ShowTags mouse right-click menu, BUT it's necessary to leave it alone, no need whatsoever to have to rearrange it every time, there must be a global variable in which your option is stored until further notice, when you optioned for "tree order", it must be tree order when you do that "F12 ab cd Enter", here it's SPEED that matters.
i ) AND NOW SOMETHING IMPORTANT: The currently available ShowTags list has nothing to do in that left pane, its natural place is the command pane! Why? Since the list of available tags for the current item is a COMMAND thing, it's for triggering views, and you are interested in contemplating these views; most of the time you are not interested in deeply contemplating the tags list, but it's the lists of relevant ITEMS you want to contemplate. So having to do ShowTags when you are in your list, just for having REPLACED your list by those tags, is sort of a "cognitive shock" every time, since, as I said, your mind does NOT want to replace your items' list by a tags' list, but wants to USE a tags' list to replace, indeed, your items' list by ANOTHER ITEMS' list! (Please note that I never ever would say something like "I'd like better" - I give reasons, and reasons that apply to most serious workers. You see, the CEO of another program - askSam - said, I use it for doing my recipies collection. Oh well, that's tremendous, but that's not serious work. When doing serious work, my hints apply, be assured of that.) (By the way, the Calendar needs to be shuffled into the command pane as well, and for the same reasons.)
j ) And even, I am asking myself at this time, should the Filter View be shuffled into the command pane as well? Which would leave the tree, all alone, in its pane. I don't know yet. If screen real estate was abundant, I would say, filter view needs its OWN pane, and perhaps that would be a solution, leaving filter view an optional thing where users could put it into the command pane when there isn't room for more, and into its own pane when working on their big screen, in their office.
Let me explain. I am formal, 100 per cent sure, about i ); as I said, I am far from sure here with the filter view. Let's analyse it. You get to your tags list in order to have another filter view, or to change, from tree, to any filter view; tags list is a command to reach in a minimum of time that other view that you want to see. In this respect, filter view has its natural place in that left pane indeed, replacing the tree. So where's the problem, once tags list is relegated to command pane?
The problem is here. Whenever you double click in the filter view (=in the left pane), the filter view is replaced again by the tree... and here it starts to become interesting. SINCE THERE MUST BE A WAY TO SEE TREE AND FILTER SIMULTANEOUSLY, and then MI would display a tremendous view of your work.
Shuffling back and forth, well, that's fine to have, but the simultaneous view of tree and filter would enhance your thinking in an incomparable way. Upon that, I'm formal again, I'm 100 per cent sure. So there must be a way to realize that.
In fact, as always in life, it's complicated. First, you go to a filter (by tags, or by my quick dialog "F12 a b Enter" that I kindly ask you to implement). And then, let's say, your filter displays 30 items, or just 20. AND you are going to do REAL WORK UPON THIS. And this implies you are going to view not any item, but let's say half or two thirds of them, let's say 15 of them, or let be just 12 or 10, IN CONTEXT, i.e. IN THE TREE...
And at this time, it's becoming awful to have to switch back and forth 30, 45, 60 times, back and forth, back and forth, now you see the tree, now you see the filter, now you see the tree, now you see the filter - NO, PLEASE!
At this time, the filter view has become the command instrument to work upon the tree, it's the tree, the context, in which your (thinking and revisioning or whatever) work will be done, the filter view will have become the ACCESS INSTRUMENT to this.
So, not every time, but extremely often in real work, that filter VIEW will become a filtered access command list for the tree, and that's exactly the point where it HAS to be of different access kind, staying before your eyes, NOT being in permanent COMBAT for your attention with the TREE, it has to SERVE THE TREE (and your thinking in it) instead of COMPETING with it; it must FACILITATE your thinking in the tree, and for doing this, it must be permanently available instead of PUNCHING IT OFF every time you want to make it work for you.
So, I think I have got to explain my point, in the beginning filter view instead of tree view can be the right thing, and whenever you are in the need to switch more than 1 or 2 times between filter and tree view, they must be VISIBLE SIMULTANEOUSLY, and that simply cannot be done when they are contained in the same pane.
That's why I made up my mind. Indeed, filter view must LEAVE the left pane, leaving tree view there alone, and must be in the command pane, but with the option to assign its DEDICATED pane to it whenever possible, by user option.
And, of course, a simple click would suffice then, since being in a different pane than the tree, the filter would update the tree view every time you'd click on an item in the filter view.
AND to conclude all this, there is always the possibility to maintain the actual state of things with the filter view BUT to give users the OPTION to shuffle it into the command pane, and the option to shuffle it into a special pane.
aa ) Technically, you could do a toggle ShowFilterPane (=that special additional pane).
bb ) AND you could do a toggle ShowFilterViewInExtraPane.
cc ) And whenever the user does bb ON, the filter will NOT be shown in th left pane (and thus not replace the tree view there), but will be shown in the extra pane, depending on aa ON or OFF: in the special pane when aa ON, in the command pane when aa OFF. And the click / double click commands would trigger displaying items accordingly.
IV
There are other professions in which work could be greatly enhanced by MI as I envision it, just think of lawyers, judges and other legal professions. I know that lawyers have their office softwares, but I also know that their office management softwares are mostly not really helpful for their REAL work, for their conceptional case work - I happen to know some German lawyer's softwares, and they are good for compulsory execution, deadline monitoring, and so on, but not so good for constructing cases... they even rely upon Microsoft Word for that part of the work.
All those professions can be addressed by many ways worldwide, the web "being there" for this, there being portals and especially professional associations' portals / sites - you just have to make them an "offer", not at all a financial one, but an offer showing to them that MI will HELP their members doing their work in a better way: make it easy for them, and they'll love you.
If MI gets to be a splendid program, I am engaging myself here to tout it, for free, in Germany, in France, in German, in French (touting it in English, well, I can write in English, I'm not so strong in speaking it...). I'll telephone those magazine people, I'll explain the splendid features and the tremendous usefulness of MI to them, I'll give them the craving for trying it out. It'll be worth it - and I'll help in doing dummy "novels", dummy psychological cases, dummy legal cases, in order to SHOW interested partys spot-on what MI can do for them, and in which easy, natural way.
As we have seen here, programs like MI don't need to be shy and confine themselves to being just information containers, no: If the GUI is perfect, they are able to greatly enhance high-brow inbtellectual work for high-paid professionals... and they, as soon as they are SHOWN what such a kind of software can do for them, will BUY, and buy at rather high prices by the way...
whereas I hope that MI's price for me will never rock beyond my reach, considering my contributions to its rise.
( Pricing is another subject, but only when the product is perfect. 200-dollars-an-hour lawyers simply refuse to work with a 50-dollar software, might it be worth 1,000 dollars... well, let's spread differenciation between basic and professional version... and without slapping in the face of vintage users... but first, let's work out an MI that would be incomparable... )
MI for Writers, Psychologists...
-
Daly de Gagne
- Posts: 52
- Joined: Wed Feb 07, 2007 6:15 pm
Fred, as I noted on another thread, just now I do believe MyInfo will be much more use friendly for writers and researchers and students of all kinds when it has the ability to open multiple windows - this feature is one lacking in virtually all other information programs.
When Eric Somer was developing ADM, I advocated for this feature (as did others, I am sure), and it was added. It was one of the things which made ADM powerful in terms of simply making it easier and faster to scan information and to return to the window where the writing was happening, or to use one window for the first draft, and a second window for revisions.
During my recent foray into the Mac world, I have looked at Scrivener, and believe it is very powerful. I have not seen anything similar in the PC world. As a writer most of my needs will be met by having multiple windows, and columns, combine with tagging. That's MI.
Re tag combinations: This was a strength of MDE InfoHandler. Surfulater offers a variation of tag combinations, which is neat. And Evernote used to have a way to deal with multiple tags, but that was lost to Windows users as part of the dumbing down of the program necessitated by the redesigns enabling it to work on all platforms.
I'd like to see MI introduce an ability to show me all items with the tags A + B + C, but not A, B, or C alone. That would narrow the search focus in helpful ways.
Daly
When Eric Somer was developing ADM, I advocated for this feature (as did others, I am sure), and it was added. It was one of the things which made ADM powerful in terms of simply making it easier and faster to scan information and to return to the window where the writing was happening, or to use one window for the first draft, and a second window for revisions.
During my recent foray into the Mac world, I have looked at Scrivener, and believe it is very powerful. I have not seen anything similar in the PC world. As a writer most of my needs will be met by having multiple windows, and columns, combine with tagging. That's MI.
Re tag combinations: This was a strength of MDE InfoHandler. Surfulater offers a variation of tag combinations, which is neat. And Evernote used to have a way to deal with multiple tags, but that was lost to Windows users as part of the dumbing down of the program necessitated by the redesigns enabling it to work on all platforms.
I'd like to see MI introduce an ability to show me all items with the tags A + B + C, but not A, B, or C alone. That would narrow the search focus in helpful ways.
Daly
Fred wrote:I
Writers
In some months, SCRIVENER will be here, in the Windows world. Scrivener is a myth for Windows users, and many Apple users write novels, screenplays, stage plays with it... some Windows users have long ago bought a Mac just for being able to use that splendid software.
Why? It's functional for writers, and it's very beautiful indeed, and people LIKE to use it.
Have any (fictional) writer. He writes in "chronological" order, using a tree, an outline, and even when he breaks up chronological order of his story, he will maintain what I would like to put as the "chronological-reading-order", or in film, the "screen order", tce same on stage, the chronological succession of stage action, be there multiple flashbacks or just "chronological-order" story-telling.
In all this, he will employ his personnel, let's say th = Thomas, ma = Mary, ch = Christopher and many others. In Scrivener, in MI also, he will tag his items. The, he will have a look upon those items (=scenes, parts of scenes, alternatives of scenes, being it a screenplay or a novel, no importance) in which th, ma, ch will interact (or, perhaps, become a subject of conversation, or whatsoever, in this case he will not do a tag "th" but "th(con)" or any other).
Let's imagine his work, let's imagine it's a big screenplay, a long novel, these tags "th" will occur a hundred times or more if th is one of the leading personnel of the novel.
He will then try to fraction these th tags, as I said before, th(con) for "here people speak about Thomas", and many more: th(i) for "innver voice", th(d) for "makes a Decision" (= upon which the next 300 pages will depend), or even "th ma" in order to describe a scene in which he's not alone, but in conversation with ma.
So, we have to consider a lot of scenes in which personnel tags are combined, and perhaps, in which those combinations (of several personnel, or of something different) will constitute something new.
Let's consider other tags, our writer will want to tag subjects, themes: ha for happiness, de for despair, es for estrangement and ra for rapprochement, I am speaking here in psychological categories of course, categories affecting the interaction of the personnel.
Of course, there's also the possibility to do rather simple tags, for towns in which our heroes move, de for Detroit, lo for London, whatever.
Our writer will consider this display of tags, 100 for th = Thomas. Then, he wants to know, when th is alone with himself, is he perhaps alone with himself too many chapters in a row, what kind of interaction, with whom, could I introduce here and there, in order to capture interest of my reader.
He will, perhaps, also introduce tags for pace: let's say, p1, p2, p3, where p1 is very detailed writing, slow pacing, p2 is "normal" writing, p3 would be narrating to us 3 months in 3 pages - of course you could do this by an attribute "P" with values from 1 to 3 or 1 to 5, whatever.
I have three points to make here:
a ) It's clear from what I've said above that our writer needs immediate access to tags, and immediate access to tag combinations. Let's analyze this further down.
b ) Our writer needs this access in "chronological order", not in alphabetical one. He will need to analyze the development of his "story", any alphabetical shuffling around of all those (perhaps 100 or more) scenes will not do him any good. If he wants to have this, he can have it, by those filter commands, but 95 times out of 100, he will need to have this filtering by tree order.
c ) He needs tags combinations very often, precisely because if Thomas is in 100 of 120 scenes, looking onto 100 scenes with "th" will not be very helpful most of the time. He COULD do this by the above-mentioned "th(i)", and so on, BUT he then will see that, when he wants to see the scenes with Thomas, he must put SEVERAL tags just for this, "th", "th(i)", "th(x)" and whatever yet - awful, not practical.
So our writers needs to have combinations, looking for "th" AND for "i", here - AND EASY ACCESS to such combinations! -, or, not simple to program, he needs some sort of hierarchical tags, in which looking for "th" will look for ALL "i" combinations, th(i), th(whatever) - WITH the possibility maintained to just looking for th, without any combinations, and finally, it would be necessary to implement the possibility to look for each combination seperately.
What I am trying to explain here is not easy, since I want to say, the MI tagging function would be splendid if it was possible to ADHERE some tags to some other tags. Let me explain, it's an important point.
When you have th, ma and ch, 3 roles, and some themes, let's say lo for loneliness, an AFFILIATION of a given theme to some role - and the possibility to do that affiliation - would make the difference.
Cou could look after th and lo, but this would also give you the scenes where it's ch who's the lonely one, and in the filter list, this Christopher scene would distract, so our writer would like to have only those scenes displayed in his filter view where let's say, there's Mary, Thomas, and one of THEM is lonely, but not third persons.
At this time, you could "program" a filter with "and" and "not" that could display that; it would not be very easy (and error-prone); having the possibility to just have some sort of "th ma lo" tags - but without any lo-tags NOT concerning th and ma - would do it.
I understand this isn't easy, and should be done in a further step.
WHAT is to be done in a first step, that's easy access to (normal) tag combinations, not only "th" OR "ma", but both (or 3, or 4) - and, in any case, in tree order.
In Hollywood, they do 300 (?) films a year, including all independent productions? In the US alone, there must be 3 million would-be screenwriters that are trying with one software or another. THE MARKET IS THERE.
Of course, throwing them MI just at it is, will not make them see the possibilities they would have with MI. Marketing of MI should include some dummy novel, some dummy screenplay, where the columns are displayed when you load the program, and where dozens of tags (and other attributes) are displayed, and where, instead of real fictional text, you'd explain how the tags and other attributes are employed, how they can be helpful in writing - and of course, we would need to have this in some major languages, English, French, German, Spanish perhaps... (I could help in French and German, even a little bit in English, for free, as soon as MI can be considered having a real tough GUI, it's all smoothness of working, I am not asking here and now ANY new feature, just better implementation of the current ones.)
Then, it would be possible to present MI to many writers' associations, many writers' portals in the web, since it would be a real high-brow contender in the Windows world, and WHEN it is, MI would naturally INTEREST all those people managing such portals, magazines... and computer magazines also.
You must see that those computer magazine journalists do NOT have the time to try out a program when you send it to them. They have a look, do NOT understand the various possibilities of it, and lose interest. It's quite another scenario when they open the program, and, instead of just seeing a blank screen, in 3 or more tabs, 3 more topics will open, each with its columns opened, and let's say, one displaying a sketch of a novel, one a sketch of a psychoanalyse, and a third one some hints how to better organize your journalist's work!
THEN those people get interested, get stuck an hour or two with the program... and will be willing to do a little paper in their magazine about it! It's all in the marketing, it's all FREE, but it's a little bit marketing work first, and above all, the GUI has to be perfect to begin with. Here, I am trying to sketch exactly this foundation of MI becoming something really splendid. MI has to CHARM people into their use of it - that's what programs like Scrivener do... and technically, they appear inferior to me...
II
Psychologists
There are some 2,000 psychologists, psychiatrists, psychotherpists and psychoanalysts in Paris alone; troughout the world, there are some millions, I think.
All those people try to organize their work some way or another; most are just working with Microsoft Word, well, let's face it, not all of them do analytical work all of the time, but almost all of them do analytical work at least some time in their work with their "clients" - and then, they are quite lost.
Analytical work is establishing the life of the client. The work isn't done in chronological order, but the tree to be constructed should be in chronological order, and of course, tagging does the work in trying to see themes, recurrent errors, recurrent self-damaging, recurrent behaviors of all sorts - for just one client, estimate some 100 tags just for these, and not including any personnel tagging (which will come on top).
Now, the therapist will constantly look into his tagging system in working with this client, and he will do it so extensively that fluidity in this tagging work will decide if he considers MI or isn't interested to employ it in his work.
As for our fictional writer, he will need to have these tags in combinations, AND in chronological order, i.e. in tree order. (And my remarks above for sophisticated tag combinations apply here, but also, my remark that normal tag combinations would be perfect for the time being, if they can be easily accessed. That very special tag combination thing is NOT necessary to realize before my marketing efforts for MI, promised, let's just have normal tag combination in an easy way, and we'll be happy for the time being!)
III
On TOP of our writer's needs, our psychotherapist / -analyst will need some automatically-driven special "tagging" / stamping system by which NOT just entire items will by tagged - that should be done automatically indeed, on top of that, but first of all, there should be a stamp with the number sign and an automatically-given number, AND that number should automatically be entered as a tag, or in a special attribute field.
Of course, you understand what I am speaking of: the session number! Whenever our psychologist does write down anything, an internal MI routine (WHEN that option is put "on" of course) will look into the tags of that item, or into a special attribute field for that item, and will check if the current session number is there already; if not, it will put it there without asking (and "behind the scenes", without opening any dialog for that: without distraction the psychologist in his work.
Before the session, the psychologist would be informed that MI has opened a new session, and ideally, it would even distinguish real sessions, number sign plus a number, and in-between-the-sessions-reviews (!), where the number sign would be with the last given session number (= and not a new one), but with a special sign after the number, indicating that the psychologist has reviewed this file after that given session, but not within the session.
Problem here, the tag / entry should be numerical altogether, so as to allow for numerical filtering, "from 30th to 50th sessions" or whatever.
This way, in any given session, the psychologist would take notes in a dozen items or so, and then, in each item, above his current writing / paragraph, there would be that session stamp, and in every item with this session stamp, there would be, in addition, that session stamp in the special tagging field. So, if a subject is reviewed again and again, that special field, over time, would be filled up with a lot of session stamps, which is why you cannot use just a numerical field for it (= which would only contain ONE number, instead of a dozen or more).
The point here is, giving to the psychologist the ability to review given sessions, and there should be the possibility to trigger a special session review mode, in which, whenever he clicks on an item belonging to this given session or group of sessions, not only the item would be displayed, but the screen would scroll to that session stamp (or the first relevant session stamp) WITHIN the item.
( In a second step, it should be possible to program a function which REUNITES those stamped passages, from all the concerned items, into a new item; for example, in order to know to which passage the session stamp applies, those psychologists would be invited to just work on paragraphs, AND on several paragraphs NOT seperated by a blank line, when doing their notes; accordingly, MI then would consider all paragraphs NOT seperated by a blank line belonging to that session stamp, and would copy them into that number x session item - of course, in doing this aggregating, MI would do another stamp into each aggregated passage, "stamping" it (=doing an additional line above each passage giving the item's title from which it comes from, and doing a blank line after each passage) - all this, again, will be another step in the future, I think; at this time, for psychologists' professional use, MI just needs the stamping feature, not the aggregation feature.)
It should be noted that the session stamp number needs to be realized by a global variable - trying to realize it by the Windows clipboard would interfere with free use of ^c / ^x / ^v throughout the psychologist's work - but this goes without saying.
III
Now for something tangible since we want to enhance MI, not in the technical way - all technical is there, I said that - but GUI-wise.
a ) I put ShowTags on a function key. There, I need the mouse to click upon ONE tag, and the filter view is shown.
b ) At this time, that's in alphabetical order. As I explained, "we all" (= I, writers, psychologists - AND lawyers also!) need it in tree order... BUT not ALL people do want what we want, so there has to be a choice. Why not doing a right-click mouse menu showing the whole list in order to option for what you want at a given time (= just like the options menu in the (not so easy-to-use, but we'll enhance that another day) filter dialog) - thus, you option for alphabetical, and all that filter display will be alphabetical, up to when you put another option.
c ) I put ShowFilter on another function key, but I see that most of the time, I am going to a), click, and c) then is displayed automatically, so this c) is good to have, but not so helpful. I also put ShowTree on another function key, and indeed, that ShowTree, I use a lot, but again, when I go by ShowTags, click there, ShowFilter is displayed, then I doubleclick there, and the tree is automatically displayed. Oh yes, I forget, to get back to the filter view, I really need that ShowFilter function key!
d ) So I have the tree view, or the filter view, in that pane to the left of my screen: perfect. Perhaps be so good and give us, additionally, a TOGGLE between tree view and filter view.
e ) And since we are at it, give us another TOGGLE between that left pane (whatever its current content) and the editor pane. Since let's face it, the problem with those two commands GoForward/Backwards1Pane is that they are no toggle, and if you do use the GoForward1Pane command to toggle, is does not work but if the command pane (=searches, attributes) is NOT open.
f ) Since we are at it, there is not only necessary that (available) command ShowCommandPane (=a toggle), but we also (=also, not instead) need a command GoToCommandPane (=and open it if it's not yet open of course).
g ) And since we're at it, at this time GoForward1Pane, when in the editor pane, goes not to the tree, but - predictably so - to the command pane, BUT then, it's NOT possible to do anything there, except when you use the mouse, that is to say, "going" to the command pane, with any command (=even by going to it with the search command it's the same problem), gives the focus just to the border of the command pane, but then it should be possible, by tab for example, or even better by DownArrow, to access there the first search result, and then, by doing an Enter, to display that item, and in a general way, it should be possible to go up and down within the search results there by DownArrow and UpArrow, triggering the item display by doing Enter.
h ) As I said before, showing the filter just by clicking on ONE Tag in a ) is fine, but insufficient; easy combinations must be possible. How? By mouse clicks? By Shift- and Control- mouse clicks? Bad! Just do, in addition, a simple command ShowTags which can be put on any function or other key, and which applies automatically to the current item, like the available command ShowTags which displays a list of all available tags. Just that this command would open a tiny dialog where you put the wanted tag(s) into a one-line field, then press Enter, so it would be "F12 ab cd ef Enter" in order to get displayed the ShowFilter with the 3 tags "ab", "cd" and "ef" - I explained elsewhere why tags as short as possible are highly preferable for most of possible uses.
This would be a very important command since it would allow us do have tag combinations displayed without going thru the complicated filter dialog, in fact it would be a very handy additional BASIC version of that filter dialog. The same option as in a ) would apply, i.e. if in the options, you had chosen "alphabetical list", this command would display the concerned items in alphabetical order, and so on. Perhaps you could add the same option field into this dialogue, as is in the filter dialog, and as would be in the ShowTags mouse right-click menu, BUT it's necessary to leave it alone, no need whatsoever to have to rearrange it every time, there must be a global variable in which your option is stored until further notice, when you optioned for "tree order", it must be tree order when you do that "F12 ab cd Enter", here it's SPEED that matters.
i ) AND NOW SOMETHING IMPORTANT: The currently available ShowTags list has nothing to do in that left pane, its natural place is the command pane! Why? Since the list of available tags for the current item is a COMMAND thing, it's for triggering views, and you are interested in contemplating these views; most of the time you are not interested in deeply contemplating the tags list, but it's the lists of relevant ITEMS you want to contemplate. So having to do ShowTags when you are in your list, just for having REPLACED your list by those tags, is sort of a "cognitive shock" every time, since, as I said, your mind does NOT want to replace your items' list by a tags' list, but wants to USE a tags' list to replace, indeed, your items' list by ANOTHER ITEMS' list! (Please note that I never ever would say something like "I'd like better" - I give reasons, and reasons that apply to most serious workers. You see, the CEO of another program - askSam - said, I use it for doing my recipies collection. Oh well, that's tremendous, but that's not serious work. When doing serious work, my hints apply, be assured of that.) (By the way, the Calendar needs to be shuffled into the command pane as well, and for the same reasons.)
j ) And even, I am asking myself at this time, should the Filter View be shuffled into the command pane as well? Which would leave the tree, all alone, in its pane. I don't know yet. If screen real estate was abundant, I would say, filter view needs its OWN pane, and perhaps that would be a solution, leaving filter view an optional thing where users could put it into the command pane when there isn't room for more, and into its own pane when working on their big screen, in their office.
Let me explain. I am formal, 100 per cent sure, about i ); as I said, I am far from sure here with the filter view. Let's analyse it. You get to your tags list in order to have another filter view, or to change, from tree, to any filter view; tags list is a command to reach in a minimum of time that other view that you want to see. In this respect, filter view has its natural place in that left pane indeed, replacing the tree. So where's the problem, once tags list is relegated to command pane?
The problem is here. Whenever you double click in the filter view (=in the left pane), the filter view is replaced again by the tree... and here it starts to become interesting. SINCE THERE MUST BE A WAY TO SEE TREE AND FILTER SIMULTANEOUSLY, and then MI would display a tremendous view of your work.
Shuffling back and forth, well, that's fine to have, but the simultaneous view of tree and filter would enhance your thinking in an incomparable way. Upon that, I'm formal again, I'm 100 per cent sure. So there must be a way to realize that.
In fact, as always in life, it's complicated. First, you go to a filter (by tags, or by my quick dialog "F12 a b Enter" that I kindly ask you to implement). And then, let's say, your filter displays 30 items, or just 20. AND you are going to do REAL WORK UPON THIS. And this implies you are going to view not any item, but let's say half or two thirds of them, let's say 15 of them, or let be just 12 or 10, IN CONTEXT, i.e. IN THE TREE...
And at this time, it's becoming awful to have to switch back and forth 30, 45, 60 times, back and forth, back and forth, now you see the tree, now you see the filter, now you see the tree, now you see the filter - NO, PLEASE!
At this time, the filter view has become the command instrument to work upon the tree, it's the tree, the context, in which your (thinking and revisioning or whatever) work will be done, the filter view will have become the ACCESS INSTRUMENT to this.
So, not every time, but extremely often in real work, that filter VIEW will become a filtered access command list for the tree, and that's exactly the point where it HAS to be of different access kind, staying before your eyes, NOT being in permanent COMBAT for your attention with the TREE, it has to SERVE THE TREE (and your thinking in it) instead of COMPETING with it; it must FACILITATE your thinking in the tree, and for doing this, it must be permanently available instead of PUNCHING IT OFF every time you want to make it work for you.
So, I think I have got to explain my point, in the beginning filter view instead of tree view can be the right thing, and whenever you are in the need to switch more than 1 or 2 times between filter and tree view, they must be VISIBLE SIMULTANEOUSLY, and that simply cannot be done when they are contained in the same pane.
That's why I made up my mind. Indeed, filter view must LEAVE the left pane, leaving tree view there alone, and must be in the command pane, but with the option to assign its DEDICATED pane to it whenever possible, by user option.
And, of course, a simple click would suffice then, since being in a different pane than the tree, the filter would update the tree view every time you'd click on an item in the filter view.
AND to conclude all this, there is always the possibility to maintain the actual state of things with the filter view BUT to give users the OPTION to shuffle it into the command pane, and the option to shuffle it into a special pane.
aa ) Technically, you could do a toggle ShowFilterPane (=that special additional pane).
bb ) AND you could do a toggle ShowFilterViewInExtraPane.
cc ) And whenever the user does bb ON, the filter will NOT be shown in th left pane (and thus not replace the tree view there), but will be shown in the extra pane, depending on aa ON or OFF: in the special pane when aa ON, in the command pane when aa OFF. And the click / double click commands would trigger displaying items accordingly.
IV
There are other professions in which work could be greatly enhanced by MI as I envision it, just think of lawyers, judges and other legal professions. I know that lawyers have their office softwares, but I also know that their office management softwares are mostly not really helpful for their REAL work, for their conceptional case work - I happen to know some German lawyer's softwares, and they are good for compulsory execution, deadline monitoring, and so on, but not so good for constructing cases... they even rely upon Microsoft Word for that part of the work.
All those professions can be addressed by many ways worldwide, the web "being there" for this, there being portals and especially professional associations' portals / sites - you just have to make them an "offer", not at all a financial one, but an offer showing to them that MI will HELP their members doing their work in a better way: make it easy for them, and they'll love you.
If MI gets to be a splendid program, I am engaging myself here to tout it, for free, in Germany, in France, in German, in French (touting it in English, well, I can write in English, I'm not so strong in speaking it...). I'll telephone those magazine people, I'll explain the splendid features and the tremendous usefulness of MI to them, I'll give them the craving for trying it out. It'll be worth it - and I'll help in doing dummy "novels", dummy psychological cases, dummy legal cases, in order to SHOW interested partys spot-on what MI can do for them, and in which easy, natural way.
As we have seen here, programs like MI don't need to be shy and confine themselves to being just information containers, no: If the GUI is perfect, they are able to greatly enhance high-brow inbtellectual work for high-paid professionals... and they, as soon as they are SHOWN what such a kind of software can do for them, will BUY, and buy at rather high prices by the way...
whereas I hope that MI's price for me will never rock beyond my reach, considering my contributions to its rise.
( Pricing is another subject, but only when the product is perfect. 200-dollars-an-hour lawyers simply refuse to work with a 50-dollar software, might it be worth 1,000 dollars... well, let's spread differenciation between basic and professional version... and without slapping in the face of vintage users... but first, let's work out an MI that would be incomparable... )
Daly, you state, "I'd like to see MI introduce an ability to show me all items with the tags A + B + C, but not A, B, or C alone." - that's exactly what I said in my lengthy way, and that's one the reasons I think the tagging system needs to be overhauled in the very first place.
At this time, using the "edit filter command", you can do some "tag is= fe ad mo" where indeed only those items are shown (and items that have additional tags beyond that)... BUT, those tags must be in order, so any item in which the tags are "ad mo fe" is not included in the hit list.
Such searches can be realized by the search, but that a totally different feature, and which does not put your items in that valuable filter list upon which you then can proceed; have a look into the help file for many details (for once), there in the search you can indeed do "tag:fe tag:ad tag:mo" - even in disorder (I did not try yet parentheses). But again, searching is not filtering, and we want our filter system to be perfect (also)!
So let's face it, many programming efforts could be spared to Petko if we did not insist on his optimization of the tagging / filter system... but if instead it was possible to have the search results not in its current search results list (in all cases), but if there was a possibility to put those results, just like results from filtering, into that filtered list (but in the command pane as lengthyly explained above).
And then it would be stored searches and so on, very easy. I suppose (without having any particular knowledge on MI's architecture) that the search functionality is rather far away from the filtering functions, but then, those filtering functions could have realized by rearranging internally-done search results, and even when this is not actually the case, it should always be possible to do this on top of it, very simply: By doing a toggle "serve search results the normal way" vs. "put this list (i.e. possible only without the tops of texts of the search results) into a new filter frame list", it's basically pouring an array into that frame instead of this one, except for the intermediate step, indeed, in which for those hits / items the currently displayed attributes would be collected in those items' records, and then displayed, like any (other) filter display, so the basic programming has been done and can be re-utilized in these instances.
In any case, there is the usefulness to have not only stored searches, but stored search results, simply those hits' arrays stored into other arrays that can be displayed by choosing from a quickly-made list, and this applies to search results AND to filter results, no need to do new filters, just store the results, then allow for doing the next filtering, but having the last dozen or so of these result pane contents at your fingertips by storing them = not permanently, those arrays would be cleared with every closing of the program.
So you see, there are a lot of things to be enhanced for those famous "cosmetic" reasons, i.e. unification of displaying results; at this time, search here, filter there, and this being totally different on a graphics level, that's not beautiful even if from a functional point of view, there's nothing wrong with it: it's the "bad" - in the sense of "not yet polished, not yet unified" conception that stucks.
( I am not speaking here of FAVORITES - with them it's the same problem in a way, favorites are not as easily accessible as they should be, and that's an important point for every new user. )
But my main point is, putting together things so that MI will be a "beautiful" program in its use, by smoothen the interaction of all its parts - that will be the first thing that should be done. There a word for it:
INTUITIVENESS, natural, intuitive use... in order to create, another big word, FLOW.
And Macs show flow, day in, day out, that's why people love them so, Scrivener being one of the most beautiful examples of that quality... but would we dare to use Scrivener for storage of 100,000... 300,000 items?
MI lets us do so, and so I'm interested in making out of MI something not LIKE Scrivener - I prove it here all around, I am not touting this detail of Scrivener and that, I just see what's HERE, in MI, and how can it be RE-ARRANGED in a way... not at all to make it a would-be Scrivener, no, something totally different, MI... but showing people, in the first 30 minutes, it's RARE BEAUTY, like Scrivener easily succeeds in doing exactly this.
To make MI FETCH people's hearts when they try it out, is my aim here. Doing Mac quality on a Windows screen, to put it bluntly.
Of course, I expected some deeper insight into my pane explanations above, but then, it's that old imagination thing, you have to have the ability to imagine something completely different WORKING before your eyes, in order to be sure if that's better, and when, why, and I know how difficult that is. I need intermediate results, beta version showing to me what already will have been done, in order to proceed with my redesigning process.
But for the panes thing here, I am positive, I am 100 p.c. sure it would do really good.
And the much-wanted "edit two items side by side" feature is extremely important, I can see that.
I'd do it this way: Doing the start command (Identify first item, e.g. ^1) for this feature, then the current item would be stored in a variable identifying it. You'd go to any other item, in the same topic / tab or any other, and you'd do a similar command (Identify Second Item, e.g. ^2), and by doing this, the tree pane would vanish, and the command pane would vanish, and your screen would be split up vertically right in the middle (subject to further rearranging, even by horizontally splitting, in which case there even would again be room for the command pane.
Above each pane would be the item's title... there would be a command to switch forth and back (toggle), but ONLY between those two panes, not going to the command pane if it's shown, so it might be the same toggle as there would be to toggle between tree and text pane, normally.
Whenever you want to go to a third pane, you'd see that the text you want to hold into the screen, has the focus, and the other will vanish, and the tree will reappear. Alterntively, you could do the tree pane as FLOATING OVER the two panes (which would be the best solution perhaps), and every choice you'll make there, will replace the item having the focus currently.
You could even preserve all the tabs above, and every time you click one tab, that floating tree pane will reappear, with that clicked tab's tree being displayed, and a special click will replace your current item (=in the text pane having focus) - it's just that you must allow for several navigations click, so perhaps you can do a toggle, and when that toggle is on, every SHIFT-click anywhere on an item's name (be it the tree or a filter view) will display that item in the currently focused text pane (the toggle itself creating the second text pane and making the tree pane (or else) flow.
Finally, it's simple. One start command (toggle) displaying the second text pane (empty up to then, or even showing a second instance of the current item by default? But that would complicate things.) - The same command to leave this special mode and automatically closing the second pane, "second" pane always meaning the pane that has NOT the focus of the two (or that has not had it last time one of them had it, it's a simple question of global variable).
And when IN that special mode, the commands ShowTree or ShowCommandPane or whatever would show them, but every Shift-click on any item anywhere would close those panes in order to preserve screen real estate for the two text panes, the ShowX commands being toggles, to as to freely see them in order to look something up; it's the Shift-click that would make the difference, no shift-click, no displaying items = no replacing items in one of the two text panes. Voilà, simple. If I overlooked something, please tell me, but you see, being a bad programmers, I always come back to simple solutions after thinking, those simple solutions being real steals for a real, and a formdiable programmer, like Petko...
and it's the way to have multiple big enhancements in a very short time, and that's what we want.
One remark about clones. In UR clones shuffle with them all their subordinate items, and hold them updated. Very very good, even necessary to use clones in a real useful way. But not simple to program, error-prone, you'd have to program quite some control routines that do a lot of checking, and so on. Well, UR does all that, these people are big workers - and there are several of them. Petko has to come up with something less convoluted, something that doesn't take him weeks just for one little function. And that's why, for example, having all my life being interested in clones, no I am interested in tags: With really good tagging, I can do without clones, especially with crippled ones. That's life. Put it simple, but simple AND ELEGANT AND FUNCTIONAL, and then, well, simple is best, and to h*** with all that most sophisticated programming helluva those UR people are able of. It's here and here only where our music plays.
Oh yes, there's also the find-and-replace thing. I hope it will differenciate between "tree only" / "text only" / "both", on top of other differenciations. (And, please, enabling replacing those ugly blanks on line ends
And one day we'll need an automatic numbering function (and even RE-numbering function) for numeric attributes.
You see, there so much to do, if we ask for really complicated things in top of that amount of work, Petko will lose his temper with us - and rightly so!
By the way, the "short tag find" ("Command a b c Enter Done") should indeed, some day, find combinations also, by "a [=and] (b [=or] c)" or whatever.
You see, it's literally endless, never ending work. Thus, putting search results into the filter pane seems a good idea to have intermediate and beautiful results without too much fuss. Consolidated beautiful results in 6 months will make us much more happy than dreams of perfect results in 10 years, right?
And so much for my love for MI for this week! ^a, ^c...
And again, text's lost, so there's need to be prepared! (I must have a look into the "User Control Panel" - perhaps I can turn the 1 hour's being thrown out into 2... or write shorter texts?????????????)
At this time, using the "edit filter command", you can do some "tag is= fe ad mo" where indeed only those items are shown (and items that have additional tags beyond that)... BUT, those tags must be in order, so any item in which the tags are "ad mo fe" is not included in the hit list.
Such searches can be realized by the search, but that a totally different feature, and which does not put your items in that valuable filter list upon which you then can proceed; have a look into the help file for many details (for once), there in the search you can indeed do "tag:fe tag:ad tag:mo" - even in disorder (I did not try yet parentheses). But again, searching is not filtering, and we want our filter system to be perfect (also)!
So let's face it, many programming efforts could be spared to Petko if we did not insist on his optimization of the tagging / filter system... but if instead it was possible to have the search results not in its current search results list (in all cases), but if there was a possibility to put those results, just like results from filtering, into that filtered list (but in the command pane as lengthyly explained above).
And then it would be stored searches and so on, very easy. I suppose (without having any particular knowledge on MI's architecture) that the search functionality is rather far away from the filtering functions, but then, those filtering functions could have realized by rearranging internally-done search results, and even when this is not actually the case, it should always be possible to do this on top of it, very simply: By doing a toggle "serve search results the normal way" vs. "put this list (i.e. possible only without the tops of texts of the search results) into a new filter frame list", it's basically pouring an array into that frame instead of this one, except for the intermediate step, indeed, in which for those hits / items the currently displayed attributes would be collected in those items' records, and then displayed, like any (other) filter display, so the basic programming has been done and can be re-utilized in these instances.
In any case, there is the usefulness to have not only stored searches, but stored search results, simply those hits' arrays stored into other arrays that can be displayed by choosing from a quickly-made list, and this applies to search results AND to filter results, no need to do new filters, just store the results, then allow for doing the next filtering, but having the last dozen or so of these result pane contents at your fingertips by storing them = not permanently, those arrays would be cleared with every closing of the program.
So you see, there are a lot of things to be enhanced for those famous "cosmetic" reasons, i.e. unification of displaying results; at this time, search here, filter there, and this being totally different on a graphics level, that's not beautiful even if from a functional point of view, there's nothing wrong with it: it's the "bad" - in the sense of "not yet polished, not yet unified" conception that stucks.
( I am not speaking here of FAVORITES - with them it's the same problem in a way, favorites are not as easily accessible as they should be, and that's an important point for every new user. )
But my main point is, putting together things so that MI will be a "beautiful" program in its use, by smoothen the interaction of all its parts - that will be the first thing that should be done. There a word for it:
INTUITIVENESS, natural, intuitive use... in order to create, another big word, FLOW.
And Macs show flow, day in, day out, that's why people love them so, Scrivener being one of the most beautiful examples of that quality... but would we dare to use Scrivener for storage of 100,000... 300,000 items?
MI lets us do so, and so I'm interested in making out of MI something not LIKE Scrivener - I prove it here all around, I am not touting this detail of Scrivener and that, I just see what's HERE, in MI, and how can it be RE-ARRANGED in a way... not at all to make it a would-be Scrivener, no, something totally different, MI... but showing people, in the first 30 minutes, it's RARE BEAUTY, like Scrivener easily succeeds in doing exactly this.
To make MI FETCH people's hearts when they try it out, is my aim here. Doing Mac quality on a Windows screen, to put it bluntly.
Of course, I expected some deeper insight into my pane explanations above, but then, it's that old imagination thing, you have to have the ability to imagine something completely different WORKING before your eyes, in order to be sure if that's better, and when, why, and I know how difficult that is. I need intermediate results, beta version showing to me what already will have been done, in order to proceed with my redesigning process.
But for the panes thing here, I am positive, I am 100 p.c. sure it would do really good.
And the much-wanted "edit two items side by side" feature is extremely important, I can see that.
I'd do it this way: Doing the start command (Identify first item, e.g. ^1) for this feature, then the current item would be stored in a variable identifying it. You'd go to any other item, in the same topic / tab or any other, and you'd do a similar command (Identify Second Item, e.g. ^2), and by doing this, the tree pane would vanish, and the command pane would vanish, and your screen would be split up vertically right in the middle (subject to further rearranging, even by horizontally splitting, in which case there even would again be room for the command pane.
Above each pane would be the item's title... there would be a command to switch forth and back (toggle), but ONLY between those two panes, not going to the command pane if it's shown, so it might be the same toggle as there would be to toggle between tree and text pane, normally.
Whenever you want to go to a third pane, you'd see that the text you want to hold into the screen, has the focus, and the other will vanish, and the tree will reappear. Alterntively, you could do the tree pane as FLOATING OVER the two panes (which would be the best solution perhaps), and every choice you'll make there, will replace the item having the focus currently.
You could even preserve all the tabs above, and every time you click one tab, that floating tree pane will reappear, with that clicked tab's tree being displayed, and a special click will replace your current item (=in the text pane having focus) - it's just that you must allow for several navigations click, so perhaps you can do a toggle, and when that toggle is on, every SHIFT-click anywhere on an item's name (be it the tree or a filter view) will display that item in the currently focused text pane (the toggle itself creating the second text pane and making the tree pane (or else) flow.
Finally, it's simple. One start command (toggle) displaying the second text pane (empty up to then, or even showing a second instance of the current item by default? But that would complicate things.) - The same command to leave this special mode and automatically closing the second pane, "second" pane always meaning the pane that has NOT the focus of the two (or that has not had it last time one of them had it, it's a simple question of global variable).
And when IN that special mode, the commands ShowTree or ShowCommandPane or whatever would show them, but every Shift-click on any item anywhere would close those panes in order to preserve screen real estate for the two text panes, the ShowX commands being toggles, to as to freely see them in order to look something up; it's the Shift-click that would make the difference, no shift-click, no displaying items = no replacing items in one of the two text panes. Voilà, simple. If I overlooked something, please tell me, but you see, being a bad programmers, I always come back to simple solutions after thinking, those simple solutions being real steals for a real, and a formdiable programmer, like Petko...
and it's the way to have multiple big enhancements in a very short time, and that's what we want.
One remark about clones. In UR clones shuffle with them all their subordinate items, and hold them updated. Very very good, even necessary to use clones in a real useful way. But not simple to program, error-prone, you'd have to program quite some control routines that do a lot of checking, and so on. Well, UR does all that, these people are big workers - and there are several of them. Petko has to come up with something less convoluted, something that doesn't take him weeks just for one little function. And that's why, for example, having all my life being interested in clones, no I am interested in tags: With really good tagging, I can do without clones, especially with crippled ones. That's life. Put it simple, but simple AND ELEGANT AND FUNCTIONAL, and then, well, simple is best, and to h*** with all that most sophisticated programming helluva those UR people are able of. It's here and here only where our music plays.
Oh yes, there's also the find-and-replace thing. I hope it will differenciate between "tree only" / "text only" / "both", on top of other differenciations. (And, please, enabling replacing those ugly blanks on line ends
And one day we'll need an automatic numbering function (and even RE-numbering function) for numeric attributes.
You see, there so much to do, if we ask for really complicated things in top of that amount of work, Petko will lose his temper with us - and rightly so!
By the way, the "short tag find" ("Command a b c Enter Done") should indeed, some day, find combinations also, by "a [=and] (b [=or] c)" or whatever.
You see, it's literally endless, never ending work. Thus, putting search results into the filter pane seems a good idea to have intermediate and beautiful results without too much fuss. Consolidated beautiful results in 6 months will make us much more happy than dreams of perfect results in 10 years, right?
And so much for my love for MI for this week! ^a, ^c...
And again, text's lost, so there's need to be prepared! (I must have a look into the "User Control Panel" - perhaps I can turn the 1 hour's being thrown out into 2... or write shorter texts?????????????)
I confess this discussion is getting way over my head, but then I remind myself that the topic is "MI for Writers . . ." -- and I am a writer. So let me offer a few non-techie observations.
I use the filtering menu in MI occasionally, and I can see that it's extremely powerful, but I also find it intimidating and difficult. As much as possible, I use the search box instead.
Here is an example of how I achieve reasonable results with the search box. I have an MI file entitled "Bibliography" that contains all my bibliographical references for various projects, both literary and personal. When I want to see a list of books and articles about a publisher/bookseller named Pickering (the subject of a book I'm writing) that I have not yet examined in the Library of Congress (which is in my neighborhood), I invoke the following macro in the search box: "custom:LC custom:no tag:Pickering". In other words, I am searching for the tag "Pickering" and the attributes "LC" (under the heading of "Source") and "no" (under the heading of "Done"). If I want, I can then add another attribute or tag (such as "Priority") or a word in the notes (such as "ordered").
This works fairly smoothly for me, but I do have to use macros for some of more complex searches, because I find it difficult sometimes to remember the names of all the tags and attributes. Evernote (sorry to bring up that alarming name again) has come up with a clever way of displaying all these elements of a complex search, but it takes up quite a bit of screen real estate.
I use the filtering menu in MI occasionally, and I can see that it's extremely powerful, but I also find it intimidating and difficult. As much as possible, I use the search box instead.
Here is an example of how I achieve reasonable results with the search box. I have an MI file entitled "Bibliography" that contains all my bibliographical references for various projects, both literary and personal. When I want to see a list of books and articles about a publisher/bookseller named Pickering (the subject of a book I'm writing) that I have not yet examined in the Library of Congress (which is in my neighborhood), I invoke the following macro in the search box: "custom:LC custom:no tag:Pickering". In other words, I am searching for the tag "Pickering" and the attributes "LC" (under the heading of "Source") and "no" (under the heading of "Done"). If I want, I can then add another attribute or tag (such as "Priority") or a word in the notes (such as "ordered").
This works fairly smoothly for me, but I do have to use macros for some of more complex searches, because I find it difficult sometimes to remember the names of all the tags and attributes. Evernote (sorry to bring up that alarming name again) has come up with a clever way of displaying all these elements of a complex search, but it takes up quite a bit of screen real estate.
Bill
I
Good evening, wsp, splendid post. I am jumping on it, not in order to be right, but in sheer delightment of your giving us a perfect example why all my lengthy posts here about consolidation of MI's functionality is indeed to the spot - since I said:
- in order for the filter dialog to be functional, you need macros, or you even avoid this dialog - that's what you do!
- that filter list is pretty and can be worked upon (=once it's constituted) for various functionality (including exporting and such... exporting to MS Word being important for sharing things, that's an important point)
- all this can be done by the search dialogue (I gave the tag example, and in the help file there are numerous other possibilities indeed, very good, that's why this help file is really important to look into a lot)
- and, there might be the possibility to use the search function BUT with getting its results poured into the filter framework, just as your search results were results from filtering
- and indeed, I spoke of stored searches, of the two flavors of them, stored results, and stored commands (without going into detail, and in fact, I'm one of those who permanently trigger macro commands in order to have those wanted features in a lesser... but then, immediate, way!
- you say it, all this fine filtering is yet to be greatly enhanced before it will really appeal to MI power users... let alone people new to it who are asked to pay, let's say 249 dollars after their trial, everything has really to be smoothend out for them...
And then, going the searching by macros detour will make place to much more filtering... or will be enhanced those ways I recommended here, in order to have real smooth "searching", be it real searching or filtering, no importance.
I am really glad you said it - you use elements intented for filtering, in search, avoiding filtering. This shows us that I am NOT - really, I do not think anybody here is really thinking this, but it's all the better to make it as clear as possible - that I am NOT, I say, attacking MI's features in order to "prove" it's not as good as that, but quite on the contrary, in the real wanting to optimize it in any possible way, in order to have something outstanding !
II
Sorry my mistake in my development for two simultaneous text panes above. When in my last post here I proned mouse shift-clicks for navigating up to the wanted second text to be displayed, that idea had been born when first I had in mind that up to that second text, the tree pane should be shown, being hidden when finally you'd selected your second item to be displayed - this would make some, perhaps five shift-clicks with the mouse in a row, just for getting to your second text and then automatically hiding the tree.
For a little automatism at the end, five shift-clicks before that? That's stupid and weird of course. (The idea was grounded in the necessity to show to MI, by normal clicking, when your navigation had come to and end, or better, when MI was not yet asked to display the second text pane, your navigation to that not being finished, thus those shift-click.
So you have your first text and want to display another one. Do the starting command for this special feature. Immediately, there will be a second text pane right from the normal one, and the will be shown also, of course. (But in that second text pane, will be blank, displaying the current text another time would complicate things too much: that old synch problem if you edit an identical text in both windows...)
Then you would navigate in the tree (or even in the tabs), in order for your getting to that second text to be displayed. And you can do this with both text panes, i.e. by changing focus between them, cou could freely get to some other item in the first, or in the second text pane, whatever you want and for hours. Just do a toggle "hide tree pane yes / no", and then it's up to the user if he will immediately trigger that toggle once he'll put those two items in the two text panes he wanted to display there, or he can continue to show the tree pane, fur further navigation to other items, or can freely show and hide the tree just as he likes.
Just one more detail, in order to avoid simultaneous editing of one item in both text panes, there should be a simple routine that would give an error message each time the user would display, inadvertantly, the same item in both panes. Of course, all navigating in the tree would not be made as in my first (bad) idea (=that was for leaving alone the first item in the then only text pane), behind the scenes that is, but simple by showing those (perhaps intermediate) items in that one of the two text panes to which you gave the focus and in which finally the target item will be displayed. And after then, perhaps, the user would hide the tree "manually" (=1 key or key combination), in order to have enough room to edit, an when he needs to navigate to another item, he'd trigger that tree showing-hiding toggle again.
THAT would be elegant and helpful (and easy to program) - not doing 5 shiftclicks in order for telling MI "I'm not ready yet". Sorry for my giving rubbish above there; my solution here will be without fault I hope!
(and now, ^a ^c, no, it hasn't been an our yet... I'll do the control-commands, you never now, do you?) done
Good evening, wsp, splendid post. I am jumping on it, not in order to be right, but in sheer delightment of your giving us a perfect example why all my lengthy posts here about consolidation of MI's functionality is indeed to the spot - since I said:
- in order for the filter dialog to be functional, you need macros, or you even avoid this dialog - that's what you do!
- that filter list is pretty and can be worked upon (=once it's constituted) for various functionality (including exporting and such... exporting to MS Word being important for sharing things, that's an important point)
- all this can be done by the search dialogue (I gave the tag example, and in the help file there are numerous other possibilities indeed, very good, that's why this help file is really important to look into a lot)
- and, there might be the possibility to use the search function BUT with getting its results poured into the filter framework, just as your search results were results from filtering
- and indeed, I spoke of stored searches, of the two flavors of them, stored results, and stored commands (without going into detail, and in fact, I'm one of those who permanently trigger macro commands in order to have those wanted features in a lesser... but then, immediate, way!
- you say it, all this fine filtering is yet to be greatly enhanced before it will really appeal to MI power users... let alone people new to it who are asked to pay, let's say 249 dollars after their trial, everything has really to be smoothend out for them...
And then, going the searching by macros detour will make place to much more filtering... or will be enhanced those ways I recommended here, in order to have real smooth "searching", be it real searching or filtering, no importance.
I am really glad you said it - you use elements intented for filtering, in search, avoiding filtering. This shows us that I am NOT - really, I do not think anybody here is really thinking this, but it's all the better to make it as clear as possible - that I am NOT, I say, attacking MI's features in order to "prove" it's not as good as that, but quite on the contrary, in the real wanting to optimize it in any possible way, in order to have something outstanding !
II
Sorry my mistake in my development for two simultaneous text panes above. When in my last post here I proned mouse shift-clicks for navigating up to the wanted second text to be displayed, that idea had been born when first I had in mind that up to that second text, the tree pane should be shown, being hidden when finally you'd selected your second item to be displayed - this would make some, perhaps five shift-clicks with the mouse in a row, just for getting to your second text and then automatically hiding the tree.
For a little automatism at the end, five shift-clicks before that? That's stupid and weird of course. (The idea was grounded in the necessity to show to MI, by normal clicking, when your navigation had come to and end, or better, when MI was not yet asked to display the second text pane, your navigation to that not being finished, thus those shift-click.
So you have your first text and want to display another one. Do the starting command for this special feature. Immediately, there will be a second text pane right from the normal one, and the will be shown also, of course. (But in that second text pane, will be blank, displaying the current text another time would complicate things too much: that old synch problem if you edit an identical text in both windows...)
Then you would navigate in the tree (or even in the tabs), in order for your getting to that second text to be displayed. And you can do this with both text panes, i.e. by changing focus between them, cou could freely get to some other item in the first, or in the second text pane, whatever you want and for hours. Just do a toggle "hide tree pane yes / no", and then it's up to the user if he will immediately trigger that toggle once he'll put those two items in the two text panes he wanted to display there, or he can continue to show the tree pane, fur further navigation to other items, or can freely show and hide the tree just as he likes.
Just one more detail, in order to avoid simultaneous editing of one item in both text panes, there should be a simple routine that would give an error message each time the user would display, inadvertantly, the same item in both panes. Of course, all navigating in the tree would not be made as in my first (bad) idea (=that was for leaving alone the first item in the then only text pane), behind the scenes that is, but simple by showing those (perhaps intermediate) items in that one of the two text panes to which you gave the focus and in which finally the target item will be displayed. And after then, perhaps, the user would hide the tree "manually" (=1 key or key combination), in order to have enough room to edit, an when he needs to navigate to another item, he'd trigger that tree showing-hiding toggle again.
THAT would be elegant and helpful (and easy to program) - not doing 5 shiftclicks in order for telling MI "I'm not ready yet". Sorry for my giving rubbish above there; my solution here will be without fault I hope!
(and now, ^a ^c, no, it hasn't been an our yet... I'll do the control-commands, you never now, do you?)
doneFred, I really like your ideas for the changes in the layout of the application and especially the reasoning behind them. Of course, I have to consider them from a technical point of view, but I think that some changes will be very beneficial for our users.
Petko, I'm delighted by your willingness to consider my ideas. This will give me others, and as stated elsewhere today, I'm not going to leave. The more I am using MI, the more I'm disappointed - yes, I put it bluntly -, but then, we have a win-win situation. I want perfect software, you are able to write it for us... and the market response will reward you, I'm sure. It's almost incomprehensible so many little things went unattended up to now, but then, your user numbers are tiny, whereas Microsoft has millions of users to give their comments on those many little faults there - with a broader market base, chances are more little faults will be reported. My remark to Karl, some minutes ago, shows my core philosophy. We need ONE program which is really good, and which is able to REPLACE a lot of other programs, we do NOT need "features". All my comments here are to be seen in light of this philosophy. And don't be afraid of Scrivener, it's a really howling thing, but it's NOT adapted yet to various businesses. MI isn't, either, yet, but it would be rather easy to do so, and then people will tout MI, like they tout Scrivener now.