Tuesday, January 31, 2006

Things about GNOME that Rock, Number Two

One comment on a previous post was along the lines of "Why aren't you praising GNOME software? The software should be number one, community is secondary." I am trying in this series "Things about GNOME that {Rock|Suck}" series to avoid talking about specific applications. IMHO that sort of feedback belongs in bugzilla, whereas I am trying to comment on GNOME as a desktop environment and developer platform, i.e. as a whole rather than particular components applications. I hope that makes sense. So, without further ado, Rock Number Two:

Increasing integration of applications, for example Evolution and Gaim. (And now Evolution and Tomboy!) The more that central information stores such as evolution-data-server (and the MIME tye system, DBUS etc.) permeates the desktop, the cooler things will be for a humble user like myself. That is real usability (as opposed to discoverability). Fewer mouse-clicks and keystrokes == more productivity in my book.

Someone else commented that I need screenshots in these posts ("people love bling"), and I was going to include some in this post. However I am currently running the latest CVS build of GNOME, and neither Evolution or Gaim work for me. I thought it was more important to test and report bugs than produce spiffy blog entries, but hey, that's the sort of guy I am...


error27 said...

Naming specific applications is more interesting than vague stuff.

It's good to list evolution and gaim but what is integrated between those two? As many specific things as possible is best.

AppleCrow said...

I'm happy to read your comments.
Because it makes people think about "simplicity & usability".

It's really sad to read that nowadays some people (which use and develop free software) think again that software is more important than user.

As professionnal user of GNOME desktop, I'm satisfied, because it increases my "speed for browse stored informations and create files" approximatively from 50%, regarding Microsoft Windows(TM).

Linus Torvald criticized the way in which the gnome deveopers' community writes the technical conditions' schedule of its applications, in "thinking for the user".

But does community have the tools to outperformed a analyze of returns and expectations from users in term of simplicity and usability?

tm said...

Keep up the good work! I see you more as an envisioneer than criticist (did i spell those two right?). Also your criticism is on point, and i think will help gnomers, when they will need references for arguments!
Soo - keep up the good work!

rob said...

I think your series is very interesting.

With regard to this specific post, heh, one could put that under "suck" as well i guess because much can be done to make integration far better (e.g. with bluetooth mobile phones).

Keep it up!
- Rob

Anonymous said...

Check if your evo bug is related to bug 329344:

(there's a patch there you can apply to gtkhtml, it fixed it for me).

Anonymous said...

But what if you use mailclients other than Evolution? Does the same work with any application a user prefers?

Integration is a really good thing, but done for some apps only, rather than in an generic way, its going to suck.

Users should not rely on e-d-s for example to have an fully integrated desktop.

Anonymous said...

Unfortunately, it is very difficult to make anything integrate well with evolution. Alex complained that the evo/tomboy integration took way longer to implement than it should have. My single biggest complaint about evolution is that there are no bindings for dynamic languages like python or ruby. I can get into e-d-s data to get contacts and calendar stuff, but I can't externally access my email without digging into the .evolution/mail files directly.

It's really a very big problem. I want to be able to integrate support for e-mail interaction into other programs I have made myself, and the infrastructure required to do so just doesn't exist. So yeah, it is nice that we have the level of integration that we have, but evolution is really poorly coded in many respects and it isn't intrinsically conducive to integration, which will limit the potential for additional such functionality in the future.