Re: [lmi] Group premium quotes: UI

[Vadim Zeitlin]

On Thu, 22 Oct 2015 12:49:55 +0000 Greg Chicares <address@hidden> wrote:

GC> >  In principle, consistency is always good, but in this case the behaviour
GC> > of "File|New" is so exceptional that I'm not sure it should be propagated.
GC> 
GC> I think of the way Hollywood extracts positives from negative reviews:
GC>   'Noted GUI expert says it's "so exceptional"!'.

 I'm rather reminded of another very Hollywoodish habit to exaggerate
everything as it's the only way I could have become a noted GUI expert.

GC> Consistency was a strong motivation. Another was reluctance to introduce
GC> GUI elements--drop-down toolbar buttons, and cascading submenus--that have
GC> no precedent in lmi: non sunt multiplicanda entia sine necessitate.

 I'd just like to note that by taking this to a logical conclusion, lmi UI
is never going to acquire any new elements which is arguably not tenable in
long term. Today we're looking at its early 90s UI with a sort of nostalgic
fondness, but in 10 more years people might genuinely not know how to use
it any more. I think it might be a good idea to gradually introduce some
newer UI elements to prevent this from happening.

GC> Furthermore, I really want exactly one group-roster icon, and I really,
GC> really want it to appear on the top-level menu so that the menu matches the
GC> toolbar more closely.

 This requirement indeed makes the use of a popup menu item almost
unavoidable. At least for the menu command, that is, we could still have a
drop down for the toolbar which is completely functionally equivalent but,
as already mentioned, more discoverable.

GC> Here's how one defunct proprietary GUI framework did it:
GC> 
GC>   http://www.faisoncomputing.com/publications/articles/owl20.pdf
GC> | When you select the File | New command in VIEWER, OWL displays the
GC> | small popup menu shown in figure 3.
GC> 
GC> Given that framework's umquhile popularity, I'd suppose that thousands
GC> of boutique applications (like lmi) work that way, or did in the 1990s,

 Perhaps... I'm clearly not qualified to argue about umquhile UIs, as I
didn't even know this word until 5 minutes ago.

GC> > course, it is also unusual when opening a new window, but at least when
GC> > opening the first one, the main lmi window is empty
GC> 
GC> Many familiar applications like word processors and spreadsheets open
GC> an "empty" file by default, which is like a blank canvas. But there can
GC> exist no empty input file for lmi:

 Sorry, I think I led you astray with this remark. All I meant to say was
that I thought using a popup menu on empty background was more acceptable
than using it when a document was already opened. I didn't mean that lmi
should always have some open document at all.

GC> For new users who prefer the keyboard, yes, the location of the popup
GC> may seem puzzling at first, but the options are discoverable...yet not
GC> instantly so. We could teleport the mouse cursor to the vicinity of a
GC> menu, but that seems rude.

 Usually the use of WarpPointer() is indeed strongly discouraged for
exactly this reason, but I think it might be one of the rare cases when an
exception could be warranted.

GC> If you can think of a way to pop it up in the "right" place, I'd like
GC> to hear about that;

 I don't know what the right place would be, unfortunately.

GC> Or as they'd say in Hollywood:
GC>   'GNU mailing list says it's "really good"!'

 And thus the day is saved again (hmm, I wonder who might be the superhero
here).

 Regards,
VZ