Actually ...
Mar. 10th, 2003 05:00 amWhile I think about it, the most welcome thing would be interface ideas.
At the moment, it's just a couple of buttons above a textarea. Should it look like a browser only more flexible, or like a newsreader?
I suspect I know what the answer to that one is, knowing my audience...
sol.
.
At the moment, it's just a couple of buttons above a textarea. Should it look like a browser only more flexible, or like a newsreader?
I suspect I know what the answer to that one is, knowing my audience...
sol.
.
(no subject)
Date: 2003-03-09 04:54 pm (UTC)There are those journals I don't 'friend' but drop in on occasionally. I'd like to be able to bookmark them in a way.
Cha-ching!
(no subject)
Date: 2003-03-12 12:06 am (UTC)Functionality-wise, I often find that after making a comment in someone else's journal, I'm inspired to write something in my own... it would be useful to be able to turn your own comments into a draft post. Preferably only allowed for comments you make, so as not to encourage wholesale plagiarism...
Also, view-wise - particularly the ability to follow threads that you've commented in easily would be very very nifty.
I make no comments about UI, other than "it should be obvious what to do". :) Also, what're you using for gui code? Ideally something cross-platform, particularly to the ActivePerl distro would be nifty...
(no subject)
Date: 2003-03-12 07:45 pm (UTC)Which is,of course, like teaching someone to ski by saying "Go that way, very fast. If something gets in your way, turn." ;)
Prototype is Perl/Gtk, on the grounds that I think it is indeed crossplatform.
Might end up doing it in wxPython, for the same reason.
You can always tag the threads you comment in, or set the config to "automatically tag articles I have responded to.". The config flag is probably a better idea.
sol.
.
(no subject)
Date: 2003-03-12 11:51 pm (UTC)Also, it occurs to me that I want thread scoring, not tagging. I want to score posts (threads) that I've commented in; score by age (both of original post and most-recent comment); score by post originator; score by specific post (which is "tagging"); score by commentor; score by number of comments... And it all needs to be ranged, ie, I might want to score posts with 10-50 comments at +100, and posts with 50-100 comments at -100, posts that have been commented in the last 5 days at +500, posts that have been commented in by *me* at +5000...
Then, obviously, I want a "friends" page which is ordered by score. :)
No.
Date: 2003-03-13 12:19 am (UTC)But that's because I care enough about UI to make someone else do it - someone who, in fact, knows what they're doing. I very deliberately avoid jobs where I have to design interfaces, for this reason.
And, apparently, because I tend to have insufficient rabbit shaped buttons.
Scoring will be Version 2.0.
Actually, given that the only thing I've really written thus far is the configuration engine, scoring would probably be trivial to add.
Actually, a configuration engine and two text inputs for do "action" to "username" is all I have written, but with the right interface, that *is* a client ;)
sol.
.
Re: No.
Date: 2003-03-13 12:48 am (UTC)IMO, the correct approach to UI design is to sit down and think about all the tasks that the user will want to do, then to group those tasks in a logical fashion and produce walkthrough documentation for all of them. :) It's not actually terribly difficult, it's just quite tedious. Some sort of flowchart is probably fine. In the process of doing the walkthrough documentation, it becomes fairly obvious if the proposed UI is bad.
If a walkthrough path for a particular function goes on and on, it's inefficient - which is okay if it's a rarely used function, but commonly used functions need to have short walkthroughs.
If a walkthrough has too many active ungrouped choices (things you might want to poke/type into to do something) to be made at any point, then it's bad, because it's confusing to the user to have to choose from so many things at once. Obviously "too many" does vary from user to user, so configurability in this area is potentially nifty.
If a walkthrough choice leads to something that is potentially "surprising" to the user, that's bad. For example, modal things should be very obviously modal - the canonical issue with VI's interface. Or for a more obviously pathological example, a GUI file-manager that allows you to delete files just by clicking on them with an ordinary looking pointer cursor.
That's basically all the heuristics you need to do "Good UI Design".
Re: No.
Date: 2003-03-13 02:42 am (UTC)It's not actually terribly difficult, it's just quite tedious., seems to me I recall hearing some people say that about running a business ;)
It's like any other specialist area. It looks like it's not that hard. It's not that hard to do a decent job of the basics. But doing it *well*, ah, that's a whole other story.
But yes, I'll cobble together something usable for the first cut of moomin .
sol.
.
[0] As with much open source software. Functionality is important - I have to use it every day. But since I'm not relying on selling it, I don't need anyone else to understand it.
Re: No.
Date: 2003-03-16 08:36 pm (UTC)Re: No.
Date: 2003-03-16 08:49 pm (UTC)But yeah, UI is sufficiently intertwingled with coding that most geeks should be able to do a half-decent job if pushed to it.
This post, btw, brought to you by Netscape 3.04. The colours on the sidebar make the text a little hard to read, but other than that, the page renders beautifully, and oh-so-very-fast...
sol.
.
Oooh!
Date: 2003-03-10 12:28 am (UTC)Probably should be configurable, because some people aren't going to like the clutter. But I know you're going to add it to the config file of doom regardless, so it doesn't really need to be mentioned.