Linked by Thom Holwerda on Fri 13th Feb 2009 23:15 UTC
Gnome A very, very long time ago I personally advocated the inclusion of a certain feature into GNOME. We set up a poll at OSNews, which resulted in a very, very resounding "yes!" from the OSNews community - many of which are GNOME users. The feature in question was the global application menubar, which allowed the GNOME desktop to have a menu bar atop the screen similar to that of Mac OS X. The poll is long gone, the debate thread in the Bugzilla has died out, and no decision has yet been made. I wanted to know where this feature stands, and how much the developers have improved it, and I was in for a surprise.
Thread beginning with comment 349047
To view parent comment, click here.
To read all comments associated with this story, please click here.
Member since:

1. This is at best going to be hacky on any unix desktop because of the issues with compatibility you have already pointed out. At least with KDE they have a very standard menu bar system where every app will work. With GTK there is a standard way of making menu bars but many apps don't use it. You combine that with Firefox (probably the #1 most used app on the unix desktop) and and you get a messy, inconsistent cludge that will surely scare away new users and irritate old ones.

Yes, you are right in the beginning it might lead to inconsistent looking desktops, but if it is well documented, people will use it. Besides, not using standard ways of doing things is usually a bad idea. It means that you will have to handle more bugs and do more testing by yourself instead of relying on code that is used and tested by a larger amount of people.

As for scaring away new users, this doesn't seam to be a big problem to MacOS, so why should it be a problem to Gnome.

2. You can't point to your poll as indicative of anything. Your sample (as others have pointed out) is self selecting. The only way to get an accurate representation is to do a random sample by contacting at random existing gnome users. That would be hard I admit but your poll certainly has no merit as indicative of what the overall community would like.

This is one of the most reasearched area in usability, and research usually comes to the same conclusion as the poll. Read any 101 course in usability, and you will get plenty of references. Even though I can't seam to remember any studie on Gnome, it would be resonable to believe that it is highly likely that you would get the same results as e.g. for MacOS-X.

3. My intuition is that what we are hearing in your poll and in the bug only sounds significant because there is a vocal minority skewing our perception. I for one hate that interface. I think if you did manage to get a good sample of people even that would be misleading compared to what the random poll numbers would look like if you first had people use the other interface for a while (where they would see the mac way is not always best and it is especially a junky cludge on unix). My prediction is that if you did a test on a randomly selected group that the poll would in fact be overwhelmingly *against*. (However I admit that is just a guess without real polling data so I recommend not messing with something that is working just fine without data)

The interesting thing is actually not if people would vote for or against this. The interesting thing is if a global menu would provide a more effective user interface (like most research suggests). People might vote against this just because it is something new that they have never tried. Once they have hade the chance to actually try it they might change their minds. For all people that remain hating it there could be an on/off switch in gconf.

If done right, it would also have the side effect of forcing developers to make better separation between control and application model. This would lead to more maintainable code.

4. Even if people do want it and it doesn't suck (both of which I would debate) you still have consider "how valuable is it?". Is it just a nicety or would it transform the desktop in some important way?

It all depends on how it is implemented. Assume that it is implemented as a "menumanager", with a standardized interface to the applications. Then it would give room for a lot of innovation, just like we have a lot of interesting windowmanagers we could have a lot of differenet menu manager, some of wich would create global menus, others may do other useful things like providing hooks for rekording of macros, or improve accessability.

Reply Parent Score: 2