The Kopete Project celebrated its
first anniversary yesterday by
announcing
the release of Kopete 0.6. Kopete is KDE's all-purpose, modular and
extensible chat client, which currently supports the MSN Messenger, ICQ,
AIM OSCAR, Jabber and IRC protocols. A sampling of the great new features
includes sophisticated (HTML) text rendering, signing / encrypting chats
and sending SMS messages to mobile phones, as well as chat tabs, translucent
windows and web presence notification. In addition, the Kopete developers
made a number of usability improvements based on
usability studies
conducted by the KDE Usability Project.
Congratulations to all involved!
Dot Categories:
Comments
0.5 shipped with some inexcusably fundamental errors (eg empty messages instead of msn user is typing), that sent me straight back to gaim. Here's hoping the last GTK app on my desktop is eclipse after I've tried this.
This is beta software. Note the version number. 0.5. Fairly fundamental errors are not exactly unexpected.
If the world were perfect, I would be in *total* agreement with you.
But it's not. A lot of software projects in the OSS/FS/License-of-the-Week world seem to have a real problem with releasing software with a version number at or above one.
Examples:
- The stable version of GAIM is 0.59
- The stable version of WindowMaker is 0.80.2
- K3b, generally considered stable, is at 0.8
- The last stable version of Mosfet's Liquid is 0.9.5
- Qtella, stable Gnutella app, is at 0.6.1
In many cases you wont see a 'stable' release of an application, ever. Fortunately, this has gotten a lot better in recent years than it used to be. However, pointing at the version number and calling it 'pre-release' is not an ample reason to shoot down criticism for what really are pretty major bugs.
Telling the person to read information on the site, wherein it prerelease status is explained, though, *is* an adequate reason.
The version number has everything to do with the current capabilities of the software. Version 1 should be stable and have the necessary features according to the developers complete. Anything less, i.e. 0.5 can be broken and incomplete in many ways.
Again, if this were actually true across all developers, I would be able to accept that as a blanket statement. But, since once again this is not a perfect world, this isn't actually how things work.
You're absolutely right in that version 1 *should* include all of the functionality and features that the developers feel the application needs. However, that's not true. This is borne out if you take a look at most of the projects I listed, especially in the case of WindowMaker.
Version numbers are meaniless unless they are defined. Traditionally a component isn't deemed stable until 1.x release, but many software components are versioned differently. Without knowing how the software is versioned you can't make a blanket statement about it's declared stability or compatibility with other components soley on the number attached to it.
They've put some incredibly nifty features in (ie tabbing, transparency etc - i just tried it, very nice) but still haven't fixed some of the basic usability problems that make an app actually usable. Whatever the version number (shouldn't you be making that comment on /. ?) the basic features should be got right before the fancy stuff.
Frankly I'm surprised you didn't tell me it was opensource software so I should submit a fix. I filed a bug, though, and will have to go back to gaim for the moment.
> Note the version number. 0.5.
And that has to do with...?
In that case, I disagree. 0.5 was advertised as "stable", and was pretty disappointing. Another major regression was that AIM messages were printed as plain text rather than HTML, rendering kopete pretty unusable as an AIM client. Looking in the code, the function to process the HTML in AIM messages had been commented out with a note saying "do we still need this ?".
The problem was immediately reported, and while there really should have been an intermediate 0.5.1 bug-fixing release, there wasn't one.
I also work on a fairly large KDE project, I know it's hard and we've had some releases with really stupid bugs, but Kopete 0.5 really was a disappointment, and so was its subsequent management (or lack thereof).
then go get use some developers that are not on vacation all the time and hand some brain over to others (the commented code was mine and I never found out WHO goddamn commented it out) :P
cvs annotate is your friend :-).
My friend does not tell me how to use him (i.e. I have no clue about annotate) ;)
Beyond the obvious "RFTM", cervisia nicely wraps it (look up the log of a file, there's an annotate button).
0.6 though should be really bug-free and stable.
But it isn't. It crashes every time I change between IM plugins.
Noatun has a similar problem and still there's no component being pointed out as guilty (the crash goes quite deep into kde libs and qt itself).
Also, why do you have to load/unload plugins all the time, isn't that something you only do once and then just go on using the application?
kopete is fantastic !!!
Multiple protocols in one GUI is more than needed...
so everybody should congratulate.
You sound as if you haven't coded one line of GPL source yet.
I don't like people who only know how to critizise my source.
And I think kopete-developer feel the same.
So:
Say a nice word first before talking about bugs...or send a patch!
I *have* coded GPL source. Does GPL mean its better or something? If someone's code is buggy beyond belief (which I'm not saying Kopete is...I use Gaim) then its buggy beyond belief, and the devs should be told. If the bugs are really, really horrible, then saying "this is crap" seems reasonable.
> If the bugs are really, really horrible, then saying "this is crap" seems reasonable.
It all depends on the "tone", just being rude saying "it sucks" is not what I call motivation.
> inexcusably fundamental errors
INEXCUSABLE??? How on earth could they be?
Since Kopete is totally free!!
grr
I would say they were inexcusable because they made kopete unusable.
ie, it was in theory a good IM client with excellent kde integration but all the wonderful features in the world couldn't make up for the fact that it crashed every twenty minutes.
ie, the bugs it contained couldn't be excused on the grounds that the featureset was rich because they made it impossible to appreciate the excellent hard work that was put into it.
ie, inexcusable because the developers must have known that they were running before they could walk.
Don't let corporations hear you using "free" as an excuse for bug-ridden software. Is that the "excuse" you want for them not using kde?
Yeah, 0.5 was complete crap. I used it for 5 minutes, until I closed the main window (which should just hide it, but keep it running, showing the docked icon) and it crashed.
0.6 is much more stable. I've just switched over from Psi. It certainly has room for improvement, but it's usable enough for me.
i've been using it since yesterday, and sadly it's crashed five or six times (that's a lot less than it used to) after windows were closed.
:(. it's back to gaim for me.
At the moment, I use gaim for my AIM needs. I'd be more than happy to switch to a KDE client (if only for the ability to dock it so the main window is unmapped)...
Basically, what I need from my client:
*The ability to use familiar keybindings (ctrl-w to delete a word, ctrl-u to delete a line, etc). I can't get along with a client that doesn't let me do this. No, ctrl-home won't cut it. Only consistency with other apps (including minicli, mind you) will.
*Hitting enter should cause the message to be sent. Not ctrl-enter, not clicking send, but hitting enter.
*Signed off clients should not be listed at all. I don't want to see them greyed out, I want them to be non-existent.
*Adding/removing buddies should be as easy as in gaim: tabs switch between online buddies and an edit tab. Extremely handy (I don't want to have to go through a config process, or a bunch of menus just to add someone).
*I need to be able to specify autologin *from the commandline*. With gaim you can do gaim --login=screenname and it will automatically log that person in. Just having auto-login isn't enough; I don't want to be auto logged in unless I specifically say so, and I want the ability to choose which screen name gets auto logged in.
These features, at least, are keeping me with gaim. It does everything in a simple, straight-forward, and non dumbed-down method that I like. That's my only complaint with some KDE apps; they try too hard to be "newbie friendly", forgetting that some of us who use KDE actually know what we're doing.
1) Not sure how easy this would be, but global editor configuration should make this easy.
2) Simple. In a conversation window, go to Settings->Configure Shortcuts. Send message, "Return" instead of the default (yeah, I think the default is stupid too, but it's easy to change).
3) Ctrl+V in the buddylist window. Or click on the looking-glass icon. Again, stupid default, but easily changed in less than 5 seconds.
4) Different developer mindsets. I honestly don't know what the Kopete guys are thinking with the add-a-buddy wizard, but I'd rather have something a bit more straightforward myself. Still, it's one of the few remaining useless pieces of the app.
5) This one's slightly more difficult. Talk to Duncan or Nick about it in IRC or send an email to the mailinglists.
My personal beef with Kopete is that I'd like to be able to have multiple accounts per protocol, but I can't currently. This is annoying. That, and it crashes/immediately disconnects with sufficiently large buddylists (in excess of 110 buddies) on OSCAR. But these are small details, and I've got confidence that they'll get fixed, so I've moved from gaim (hopefully forever).
-clee
> (if only for the ability to dock it so the main window is unmapped)...
GAIM does this. Just enable the "dock" plugin in the GAIM preferences and it will put an icon in the kde status dock. Clicking the dock icon toggles the buddy list visibility. And right-clicking it allows you to set options without having to map the main window. YMMV, but it works for me with GAIM 0.60 and KDE 3.1
> At the moment, I use gaim for my AIM needs. I'd be more than happy to switch
> to a KDE client (if only for the ability to dock it so the main window is
> unmapped)...
You can dock *any* application by using the ksystraycmd tool that is included in kdebase. ksystraycmd --help for more details.
Rich.
Thanks for the tip.... I didn't know about this one. Makes my life MUCH easier!
You're welcome. There's not much point in writing these things if people don't use them. You might also want to play with kstart (a similar tool for customising window decor) if you get bored. :-)
Rich.
Right click on the Kopete icon in the system tray, select minimize. It's the exact same way for KMess as well. I think the guy is right, it's more of a KDE build in feature, so why code it into your program ?
* Kopete is a KDE application, and as such uses all KDE standard edit shortcuts as specified in KControl.
* Again, Ctrl+Enter is the default because it is a defacto KDE standard for "Send Message", as used by KMail, KNode. If you want to set this to just plain enter it is as simple as Settings->Configure Shortcuts, as with all KDE applications.
* As stated below, click the button in the toolbar. This preference is saved.
* This is a matter of preference. Personally, I prefer the wizard.
* This feature may be added, it seems reasonable.
I did not know how to set enter instead of control enter in KDE apps. That is a very helpful and useful thing to know. Now that I can use return to send messages, kopete is my default IM client. THank you very much.
Ah, when i set settings->shortcuts to change for sending a message from clt+enter to enter, it doesn't take affect. Enter remains as new line and clt+enter remains as send message. It ignores that setting (kopete 0.8.1)
yeah for me too... even setting as the shortcut for send it still doesnt work... and have to send by
Same in 0.09. Needs to get fixed!
Same in 0.90. Needs to get fixed!
wouldn't it be a good time for kopete to move over to kdenetwork?
Or kdepim.
perhaps once it stops crashing doing fairly rudemantary things.
I *really* hope it stablizes before KDE 3.2 tho.
I use the 0.6 Kopete on Mandrake with
MSN and ICQ protocols. It never crashes.
What are you doing with it?
Look at the attached SS:
I'm using it with AIM (oscar) only.. after a few hours, it keeps on using 95-99% CPU usage. No idea what's causing it, since this is hours after not doing ANYTHING with the app.. the app is still usable (but slow). I have no way to reproduce it even other than just letting it run, heh.
This is with 0.6, with KDE 3.1, and .. I had other major stablity problems with 0.5, but they seem to be fixed.
I love kopete (especially the interface), but it's not there yet for me :(
Also, it should have a wizard type of thing to select what protocols the person wants to use.
> I'm using it with AIM (oscar) only.. after a few hours, it keeps on using
> 95-99% CPU usage. No idea what's causing it, since this is hours after not
> doing ANYTHING with the app.. the app is still usable (but slow). I have no
> way to reproduce it even other than just letting it run, heh.
Try disabling the "Connection Status" plugin. I had the same problems (with ICQ tho) when I used it.
Thanks, but I didn't even have this plugin enabled in the first place :(
Also try disabling one of the two plugins to find out which one is guilty so we get an idea of where to search for that bug.
Things like
"it hangs"
"it sucks"
"it crashes"
etc. won't give us the opportunity to fix bugs.
there are many chrashes whenn you use IRC
It will be in kdenetwork for kde 3.2
Yes fine, but if it will be in kdenetwork for KDE 3.2, why is it still in kdenonbeta?
"translucent windows" + "usability improvements" = unconvincing.
Quite a lot of work went into API restructuring to allow better maintainability in future versions (and lay the groundwork for multiple accounts per protocol and KDE address book integration), but as that makes little sense to most users, I guess Dre left it out from on purpose.
I have no doubts that a lot of work went into this new version of Kopete. My post merely alluded to the incongruity of those two phrases being used together.