Hyper-filters: hints, wishes (especially for shortkeys to stored ones)

Discussion & Support for xplorer² professional

Moderators: fgagnon, nikos, Site Mods

Post Reply
aml
Member
Member
Posts: 43
Joined: 2012 Sep 18, 08:24

Hyper-filters: hints, wishes (especially for shortkeys to stored ones)

Post by aml »

1) The "hide-show folders" and the "hide-show-filter" toggles aren't that helpful for me since the "no filter" toggle resets the "show folders", too, since the latter setting (i.e. show folders or not, and then, how) is a global setting and does NOT work on a specific pane; I systematically need folders just in one pane, but not in the other one, and since the global folder setting is not pane-specific, as said, I must leave the global setting to "show them", but then must do anything by hyper-filters (the Alt-h dialog).

To make this more clear: In order to get folders, too, in one pane, I must leave the global folder setting alone.

> I then hide the folders in the other pane by hyper-filter (with the filter setting "without folders").

> Then a global view of the folder (by the filter-toggle, in that "no-folders-wanted" pane) will show the folders again, so that for hiding them I also need to toggle the "hide folders" toggle again. I.e. unwanted interaction of the toggles, or in other words, the "hide-folders" toggle is not persistent, for the pane in question, but reset to "show folders" when the "filter" toggle is reset to "no filter".

> Hence my need to do it all by hyper-filters.

That's not a problem per se since you simply can do an otherwise "empty" hyper-filter with just the "Folders" box UN-checked (as I do it with the other, "real" filters for that pane, while the other pane, due to the global setting, SHOWS the folders, too).

Thus, a better interaction of the toggles would be welcome, and/or a specification of the global "how to display folders" setting for left pane and right pane.


2) Here a little hint: Since in the "Named" field of the hyperfilter dialog, there is (?) NO possibility to EXCLUDE (sub-) strings, i.e. NO regex allowed in that field (compare with the "Containing text" field (i.e. the one for file-CONTAINED strings) where the possibilities are much larger), I filter by an "additional rule" which is an otherwise "empty" filter (i.e. nothing in the "Named" field) but which says "Name contains "-x*" (exact=0)" and which in fact is "Property: Name" (column) and "Text settings - Find:" -x* - this shows everything (or in my case, all files only) in the folder, except for files (or folders, according to the toggle) starting with "x".

Thus, more "direct" action onto the "Named" field, as for the file-content field, would come handy, but the missing string functionalities in there can be realized by additional rules for the "Name" column.


3) The hyper-filters are extremely powerful, AND can be stored (and how to delete them in case, is now thankfully explained in a mouse-over), and I particularly like the fact that simple browsing (!) the drop-down field for the stored filters will immediately show their details, no need to select one for showing, then select another one, etc. - this is a brilliant detail indeed.

But I miss, are assignable shortkeys for these stored filters, shortkeys that would work from anywhere within x2, and without flashing the dialog (at this time: alt-h for opening the dialog, then alt-p for "Predefined", then "x" or any other 1- or several-chars string in order to select the wanted filter, then 2 times "enter" (first to close the drop-down, then to do the filtering).

It's clear as day that such a powerful filter functionality should get the prominence it deserves, so any fiddling around with it, for standard (whilst in case, elaborate) filters you need again and again, should be avoided, their accessibility should be immediate, hence the need for x2-global shortkeys for specific stored filters: Pressing the "Save" button could then open another, "Assign shortkey" dialog, the shortkey field remaining empty, in case, OR it could be a new sub-menu within the "Customize keyboard" dialog.


4) And another hint again. Filters for (sub-) strings in the "Comment" column (i.e. for tags or other content) are very useful. When you run a (any) filter on (sub-) strings there, display of the filtered results will take quite a lot of time (depending on your storage of course), BUT "next" time, it'll be very smooth, even for NEW entries in that column. This means that x2 obviously uses some array or such in order to store it all, on first run of such a "Comment" filter, and then duly updates that specific storage with any change or addition to that column, in order to deliver immediate results further on. Thus, the very first ("trial") use of this functionality should not deter users from really using it, it can become a perfect ToDo system for example, also speed-wise, especially with shortkeys for often-used filters. (Reservation: I did not try it with multiple folders, such "Comment" strings, and switching around between those folders then, BUT I use it with one big folder with 4-digit number of files and (high) 3-digit number of files having got such (various) "comments", and filtering here is ace.)
User avatar
nikos
Site Admin
Site Admin
Posts: 15771
Joined: 2002 Feb 07, 15:57
Location: UK
Contact:

Re: Hyper-filters: hints, wishes (especially for shortkeys to stored ones)

Post by nikos »

there is NO possibility to EXCLUDE (sub-) strings, i.e. NO regex allowed in that field
you are wrong in both counts :)
you can have something like

Code: Select all

*.txt,-*test*
which would show every txt file unless those with test in their name
you can also set the numeric field next to the Named box to 0 which will enable the regexp mode

for easy access to saved hyperfilters you can use the Autofilter toolbar button, and they are all under Saved filters submenu. Sadly it isn't possible to add a keyboard shortcut to this menu, but you can define a macro that opens the filter dialog and selects a filter by name (if you have version 4), and then set a keyboard shortcut to this macro, see
www.zabkat.com/blog/menuhack-customize-toolbar.htm
aml
Member
Member
Posts: 43
Joined: 2012 Sep 18, 08:24

Re: Hyper-filters: hints, wishes (especially for shortkeys to stored ones)

Post by aml »

Thank you!

1) Numeric field = 0 > regex indeed, is stated in the mouseover of that field (not of the "Named" field); also, didn't even try "-x" since its mouseover: "accepts partial filenames and wildcards", "-" being neither; also, field "Containing Text" has option "verbatim", so I would have tried negation there in case but field "Named" has no such options, just the mouseover, and thus, by taking the (incomplete) mouseover for final, I (wrongly) deducted that except for the "wildcards", anything in that "Named" field WAS verbatim.

Of course, this greatly simplifies the needed "additional rules", and the regex possibility is great by its own count!

2) Oh yes, I had not thought of the new macros now in version 4, am on 4.

Hint: Just make a reasonable, fixed price for upgrading to "Ultimate incl. Lifetime" for ANYBODY having "Professional incl. Lifetime", and you'll get lots of such upgrades, problem for users in that case currently being that they lose the paid "lifetime" part of their (henceforward useless-to-them) "Professional" license by upgrading, so many will not do it; make 2 different, always reasonable prices for these people having bought by "bits" if you really deem there is the need to differentiate these groups, bearing in mind the "bits" "purchase" price had been reduced then, but not the "lifetime" price add-on; reasonable being price difference between Prof and Ultimate, in case between Prof-really-paid and Ultimate, plus price difference for the lifetime add-on between Prof and Ultimate; if there is no such price difference, plus some 10$ perhaps (or 20$ if you do NOT insist also on the price diff from bits-license to regular-license at the same time), but losing the "bits" price advantage AND the, now-worthless, full lifetime add-on for the now-shelved Prof. license is simply too much loss on the user's side, especially when the Ultimate version just appeared some tiny months later than the purchase of the "Prof. lifetime".

3) "but you can define a macro that opens the filter dialog and selects a filter by name" - yes, I do this currently by Autohotkey macros, but it would be oh so much more elegant, and less time-consuming, to NOT have the flickering of the opened filter dialog, but to execute the stored filter "behind the scenes", with no visual bothering the user, and immediately.

4) Since the stored filters aren't accessible but by going over that dialog currently (and be that triggered manually or by macro), your choice that the alt-h opens the dialog with the previously-used filter, makes sense, but at first sight only. On the other hand, this choice doesn't make immediately-available a generic, "empty" filter, but the user sees the dialog with all the settings of the previously-used filter, and the idea "whenever the user uses a filter, chances are they will want to use the previously-used filter" is already served by the toggle "filter on-off" (which also applies the same filter to other folders, the user possibly having switched to in-between).

So, precisely because of the existence of this toggle, chances are that whenever the user triggers the dialog again (alt-h), they will want to use ANOTHER filter than the previously-used one (i.e. any other: a stored one, or a new one), and so it would be logically correct to have the dialog empty upon display.

I helped myself with a preset, empty filter stored as "g" for "generic; alternatively, that could be empty filter "0", in order for it to appear first in list, but that would only make sense if the default filter, i.e. the one on display upon alt-h, would be the first in list, which, as said, is currently not the case.

5) For generic filters (and just ONE entry in the filename field), a global setting, or a checkbox (with persistence if checked previously or not in that generic filter), "string as start string" (i.e. x2 adds a "*") would be very helpful for many users, especially for all those who have no U.S. keyboard, and for which adding a "*" thus will mean shift-somekey.

Of course, the current default (string is substring anywhere, not necessarily at name begin) should not be replaced, there's to many users relying on the current way it works, but sparing us other users the (for us) systematically-needed shift-key "*" (by global or specific toggle, i.e. by setting or by checkbox) would be a relief indeed.
User avatar
nikos
Site Admin
Site Admin
Posts: 15771
Joined: 2002 Feb 07, 15:57
Location: UK
Contact:

Re: Hyper-filters: hints, wishes (especially for shortkeys to stored ones)

Post by nikos »

if you have pro+lifetime you can upgrade to ultimate+lifetime paying just the difference in price. Please send me an email about this conversion

as we speak I am adding a new macro option that won't show any dialogs so you don't have to see flashes
aml
Member
Member
Posts: 43
Joined: 2012 Sep 18, 08:24

Re: Hyper-filters: hints, wishes (especially for shortkeys to stored ones)

Post by aml »

1) Upgrade: Fantastic, thank you very much! (On bits, the question had arised here and then, and then you just answered to be contacted in private, i.e. on a case-by-case basis, so many people continue to fear to lose their previous investment; so this clarification helps a lot!)


2) Internal macro: I hope I can grasp how to do it, functionality is brand-new, with not too many examples up to now, but this is really fantastic in perspective!


3) The alt-h dialog again: It works in a very weird way, even if you do it all manually, i.e. not by an external macro (with perhaps too few milliseconds wait time between the actions):

Let's say you have some predefined filters, named "a something", "b something", and so on (but it's the same behavior if you just name them "a", "b", etc.). (Since there is no dedicated "comment field", just the "name field", called "predefined", i use 1-to-2-char "names" and then put a short comment within (), after a space, which technically belongs to the name, too, then. But that's not the problem.)

(Any) filter "n ..." is running, then you want to switch to (any) filter "t ..." (any name, any position in the list, it doesn't matter). You do it this way:

alt-h (opens the dialog)

alt-p (select "Predefined")

down (shows the list) (Or you leave this step out if you know the name anyway, and if there is just one named "t" and nothing with "t*", but this gets you into even more trouble, depending on the current entry, so you press "down", to be sure.)

t (you put in the first char of the name, which as said may even be the full name)

If you do "enter" now, most of the time, you will just get the previous filter ("n" in my example) again, i.e. the changes in the fields are NOT yet triggered most of the time, and that's really weird since a simple "enter" would be logical here.

Instead, you'll do a second "down", in order to be sure the wanted values display AND remain stable in their fields.

And then only you can (and must) do the "enter" which you erroneously had thought to be sufficient before. So, even after the (unambiguous) "target character(s)" input (into the correct, the "Predefined", field), instead of just pressing "enter", you must press "down, enter".

And then only you'll press enter again, in order to trigger it all ("filter" button)

Thus a simple switch from one filter to another becomes:

alt-h
alt-p
DOWN
(unequivocal) target char(s)
DOWN
ENTER
enter (trigger it all)


4) The "*" again in that dialog (inconsistency):

Above, I have said that in general, in the "Named" field, many people will need to manually supply the end-"*" (and which on their keyboard may not come handy) for their filters, so I've asked for a (global or specific) toggle "* not implied / implied", "not implied" being the default setting of course.

Thankfully, in that "Named" field, a (manual) end-"*" implies that the whole string is "xyz*", which means that it's not in the middle of a file name, but at its beginning: it's not "*xyz*" - in other words, the set "*" at the end implies that where there is no "*" (i.e. at the beginning), there is NO implied "*".

Unfortunately, it's not the same way within the "Additional Rules" field, and this applies to "include" as to "exclude":

If you want to EXclude name beginnings from there, you must do it by regex there, or again within the "Named" field instead, since the "*" in "Name contains "-xyz*"" here is redundant, a simple "Name contains "-xyz"" would do - in other words, the string "xyz" is always searched for within strings in the filename, not (only) at its beginning, and this means the rule will exclude many files/folders you would like to be displayed though.

Similar with INcluding file name beginnings in the "Additional Rules", a "Name contains "r*"", with the redundant "*", works exactly as "Name contains "r"" would.

In other words, any "xyz*" in the "Named" field works as "begins with", while any "xyz*" within "Additional Rules" (for the "Name" column at least) works as a simple "xyz" and thus in- or exludes many unwanted entries.


5) And finally, the "and/or/not" combi in the "Additional Rules" field: I miss parentheses there, and I admit that here, I did not dig deeper into the problem but gave up rather soon, my (real-world) tries being not successful, so I try now with regex instead:

- does it (the combi) include "Named" and/or "Containing Text",

- or is it applied just to the elements within the "Additional Rules" field?

And with more than 1 element in that field:

- all "and" = simple
- any (simple) "not" = simple, but

- has "or" got precedence over "and", so that "a or b and c" (3 lines) would be processed as "conditions (a or b) and c", or

- is the order important (I have read in the help that internally, x2 will try to process speed-maximized, but that doesn't necessarily affect the idea that the order in which the user puts those elements into the field, could be evaluated as (simple) parentheses), e.g. so that "conditions (a or b) and c" could be written "a and b or c" (3 lines again), for x2 processing all "and" following immediately each other as one group, and then process any "or's" after it as independent "or's"?

It's understood that such processing artistry would only help for quite basic combinations of conditions, but then, since there are up and down arrows for the user to reorder the conditions, is this just for visual means, or can the internal logic be influenced by this order? (Help and mouse-over don't elaborate on this.)

It seems the user can avoid some of the problems by combining several simple lines into one line using regex, but it's evident that for most users non-programmers, parentheses would be much simpler to use; as for their implementation, I suppose that could be rather easy as long as you only allow for simple combinations like "(a and b and c) or (c and d)", "(a or b or c) and (d) and (e or f)", not parentheses included within others?
User avatar
nikos
Site Admin
Site Admin
Posts: 15771
Joined: 2002 Feb 07, 15:57
Location: UK
Contact:

Re: Hyper-filters: hints, wishes (especially for shortkeys to stored ones)

Post by nikos »

3. with all combo boxes, the way you use them requires pressing the DOWN key to select the item whose first character you just typed. Then you will see all the filter details in the various fields. To save on keystrokes, start typing what you're after and xplorer2 will show you only the matching entries, which you can select with DOWN then ENTER

4. indeed the * is handled differently in NAMED field and if you use it within Additional rules. This is because in the latter case it is meant as matching a substring within a larger buffer (text). You have two options to match the beginning of a string

* use a regular expression ^xyz (notice the leading ^ which means beginning of string)
* use : for an exact match, e.g :exact will only match "exact" and nothing else

5. in xplorer2 there is only one level of Boolean expressions. All rules are stringed together in an AND fashion unless you make some particular rule OR or NOT. There are no brackets possible. For more flexibility have a look at deskrule that can do arbitrary Boolean expressions explained here
aml
Member
Member
Posts: 43
Joined: 2012 Sep 18, 08:24

Re: Hyper-filters: hints, wishes (especially for shortkeys to stored ones)

Post by aml »

"3. with all combo boxes, the way you use them requires pressing the DOWN key to select the item whose first character you just typed."

Of course, but we're speaking here of selected-combo-box, and then an unequivocal entering of the ONLY fitting character/string, so that no (even-partial-)list-display would be necessary, this becomes evident in the case of a simple "a" when no a* entry is there.

"Then you will see all the filter details in the various fields. To save on keystrokes, start typing what you're after and xplorer2 will show you only the matching entries, which you can select with DOWN then ENTER"

From your explanation and trying it, I now grasp that the "matching entries" you speak of are NOT only at the beginning of the stored names (as would be standard I think), but are to be anywhere in the file name, which multiplies the number of "matching entries" of course: by "a", I mean "a" in case, or "a*" if there are several "a...", you and x2 mean "*a*" by "a": That complicates things, of course, and explains the seeming unreliability of my above-described "regular" use of the dialog which I then had to replace by the described convoluted use, in order to get it reliable. (I don't remember to have encountered any other drop-down menu in my whole life where an "a" wasn't meant "entry begins with a", but "being non-standard being the standard way for x2", I should have guessed something around these lines.)


4. " :exact will only match "exact" and nothing else": Thus:
"abc" is "*abc", and
":abc" is "abc as word-only"? (and "verbatim" ":abc" would be the string ":abc" anywhere of course) - why using "exact" for "word-only" when we know what "word-only" is, but not what "exact" is, all the more so since "verbatim" is something else than "exact" here, while we probably tend to see those as synonyms? Anyway, the difference between string processing in the "Named" field and the "Additional Rules" field is worth noting since in many cases, with many entries, wrongly-constituted result lists will not immediately appear as faulty - hint in mouseover?


"5. in xplorer2 there is only one level of Boolean expressions. All rules are stringed together in an AND fashion unless you make some particular rule OR or NOT."

This is sufficient explanation for "AND" and for "NOT" ("not" being an "and not", the "not" referring to just the current element); I said above these are easy. But for "OR", we need to know to which element(s???) they refer:

a) and b) within the "Additional Rules" field only, the other fields ("Named" and "Containing Text" being connected to the "Additional Rules" field with an "AND":

a) to the ONE immediate-previous element? (i.e. user order would be able to do some "virtual parenthesing" (=precedence of "or" above "and")

b) to ALL previous elements? (them token as one parenthesis, and the "OR" line as the other? =precedence of "and" above "or") (here again, the order the user "does it" would determine the processing)

What about several "and" AND/OR "or" in the "Additional" field? There ARE (implicit) parentheses where there is "OR" AND "AND", necessarily.

("OR" above "AND" would be standard, so
AND x
OR y
AND z
would be read: (x or y) and z
not: x or (y and z)

A hint in the mouse-over would be welcome.


Between yesterday to today (new "session"), I lost predefined (and then, of course, "saved") "rules" (not many, thankfully); I remember I had deleted the "sample" (deletion was possible), so that might be an explanation, since "sample" is again there today. Could the deletion of "sample" have triggered a "reset" to the state of the day before, upon (normal) closing x2? (I let alone "sample" today but fear possible deletion of other "rules" overnight, by other causes?)
User avatar
nikos
Site Admin
Site Admin
Posts: 15771
Joined: 2002 Feb 07, 15:57
Location: UK
Contact:

Re: Hyper-filters: hints, wishes (especially for shortkeys to stored ones)

Post by nikos »

3. yes the match occurs anywhere like *abc* so you'd have to type 3-4 characters to pin point what you are after

4.:exact is only matching "exact", no words or subwords, the whole thing must be "exact" or it won't match

5. the flat Boolean expression is evaluated top to bottom so the brackets are (A OR B) OR C
there is room for confusion though if xplorer2 reorders your list, putting slow columns further down

xplorer2 shouldn't forget settings unless you had many windows open and changed settings in one of them. If you discover any reproducible sequence that shows xplorer2 forgetting please let me know
aml
Member
Member
Posts: 43
Joined: 2012 Sep 18, 08:24

Re: Hyper-filters: hints, wishes (especially for shortkeys to stored ones)

Post by aml »

3. Thank you for the confirmation. As it's totally non-standard, it'd be really helpful to have a mouse-over hint or similar, in order for the users this not discovering at their expense only, after some time.


"4.:exact is only matching "exact", no words or subwords, the whole thing must be "exact" or it won't match" - Sorry, but I don't understand a mumbling word of this, WHERE's the difference / differentiation you make between the "(whole word" vs. "partial string") term we all are acquainted with (string within a "word" or "as word", whilst "as word" being enclosed by "non-word" characters like ^r^n, tab, space, comma, dot, etc., and your ":exact"? I simply don't grasp it, and I'm certain neither 90 p.c. of readers of this thread will do?


"5. the flat Boolean expression is evaluated top to bottom so the brackets are (A OR B) OR C" - you probably meant (A or B) AND C, and also (A and B) or C, i.e. you obviously wanted to express that there is NO implicit "OR over AND precedence", and thus NO grouping either, so a mouse-over hint stating this would be helpful. As for the speed-optimizing, internal, possible re-arrangement of the lines in that box, we've got a real problem here, since thus, by stating the order of conditions, the user can NOT implement implicit parentheses, but must expect unpredictable outcomes, which is not acceptable. Hence my suggestion to implement explicit (simple, 1-level-only) parentheses, instead of your currently "top-down, single-line-by-single-line" processing, unfortunately scrambled by subsequent re-arranging the lines for speed reasons but which possibly breaks the implicit parentheses the user might have had in mind.

The current situation is not acceptable; if you decide to not implement explicit parentheses (for now), please delete the re-arranging-for-speed-reasons of the "lines" anyway; whilst the idea behind speed-optimization is perfectly understandable on first sight, any creation of unpredictability by those means are unacceptable, unfortunately: the user MUST know what a certain input of them will produce as the result.

I understand that your striving for speed-optimization is motivated by what database engines do. BUT they preserve the (explicit or implicit) logical parentheses by doing that, whilst your code doesn't (and cannot) do this, so please refrain from that try: There are just a handful of SQL db's worldwide, and there's a reason to it. Especially here, but with general scope: If really necessary, my previous post re "Columns" some minutes ago will have proven, I think, that I consider x2 being THE foremost file manager today, and so please believe me that I just want it to be perfected by you, Nikos: just leading a mediocre pack isn't enough by my (conceptual, not programmatical) standards, let's strive for real superiority, nay? (As for your - impressive indeed! - query software, it's - it'll be! - all about its integration into your file manager (the "Ultimate" version, of course!).)


Lost "predefined" "additional rules": No further losses between yesterday and today; my x2 is always "just one of them allowed" (yes, I have taken note of the/your alleged difference between "instance" and "just more additional windows" while NOT grasping it really; I always just have ONE x2 window open, and my macros always revert to that one, single x2 window, not triggering a second x2 instance); I always suspect my deletion of the "sample" "additional rule" entry before yesterday, without being certain of the possible cause of the losses between the day before and yesterday, as said).


In general, whenever you divert from the standards, i.e. from the way users think things will work, from their experience, at least "say so", and we've discovered, just within that minor alt-h subset of the functionality of your brilliant file manager, lots of non-standard ways which have not been communicated (neither by the help file nor by, most needed-then, I think, mouse-overs) yet to the "lost" users: This harms your business, and multiplies remarks all over the web re x2 in the line of "technically really, really good, but terrible user-interaction", while - NOT superior in any way - DO multiplies sales, by almost incredible buzz-marketing which obviously is as unjustified as Apple's one.

I perfectly understand your striving at "doing things better than they're ordinarily done"; it's just the necessity to also:
- communicate those derivations very clearly, and yes, also: to
- provide the ordinary way to do things, optionally, whenever the user in question can not follow your "special" idea but wants to continue their file administration the "traditional" way, BUT with some superior goodies only x2 can deliver: it's all about striking the perfect balance, and since programming makes available options by nature, that'll be easily attainable for the superior coder and software designer that you are, Nikos!
User avatar
nikos
Site Admin
Site Admin
Posts: 15771
Joined: 2002 Feb 07, 15:57
Location: UK
Contact:

Re: Hyper-filters: hints, wishes (especially for shortkeys to stored ones)

Post by nikos »

4. exact means exactly the lot. Not words, beginnings of words or what have you. It's either all or nothing. For example if a property is "hello world", with the exact notation you would have to search for :hello world. Without the : you could match hello or world or "wor" separately

5. sorry but if you need exact specification of Boolean rules you'll need to get deskrule, which is sold at a discount for xplorer2 owners. For most people where all additional rules are AND, brackets won't make a difference
aml
Member
Member
Posts: 43
Joined: 2012 Sep 18, 08:24

Re: Hyper-filters: hints, wishes (especially for shortkeys to stored ones)

Post by aml »

4. Thank you for the explanation, so you use ":string string..." for what's generally called a "phrase search": "string string..." (WITH the ""); I would never have guessed that! Of course, that (in your example), some regular "abc xyz" would also find "x" is news that users should know indeed; at the most, a hit for "abc" or "xyz" (implicit "OR" by the space) would have been guessable. (I personally try now to do it all within the "Named" field, by regex, in order to avoid the problems I've discussed above; except for attributes that is.)

5. DeskRule is one of a kind indeed; but as said (or at least implied by me), I suppose most users would like to have the complete avaivalable functionality within their one-and-only (and accordingly spiced-up, and accordingly priced-up) file manager, instead of having to switch forth and back between two different programs for special functionality, especially when both programs come from the same (brilliant) developer.


I'm not advocating multi-line mouse-overs per se, but if the relevant info cannot (or cannot easily) be found within the help pages (because they are probably too long, because of being multi-subject, in some instances, or because the info simply isn't there), more explicit, even multi-line, mouse-overs are of big help; ditto, btw, for any info which is available from the help, but which is probably not counterintuitive, and where the user thus risks, from one rare use, to the next one, months later, to have forgotten the deviation from the usual way, their memory providing wrong results how to do it. I admit this risk is higher with types of software like file managers where users often tend to use several pieces of software upon the same data, instead of just using the same software again and again, so that get really intimate with those deviations from the respective usual way, over time, and thus don't put a strain on the user's memory anymore.
Post Reply