Across the various planets, there seems to have been quite a lot of discussion about interfaces and how Gnome 3.0 should proceed. I think the most interesting of all these posts has been this one from Alberto Ruiz where he linked to this talk given by Aza Raskin. While I know very little about interface design and human computer interaction, it seems to me that many of his suggestions make a lot of sense and could help to tackle some of the problems people have with their computing experiences. I’m going to share some of my thoughts on this here, take it or leave it, but I at least hope those who are working on Gnome 3.0 interfaces will give some thought to Aza and his father’s work.

Ubiquitous Search

One of the major points under consideration for Gnome 3.0 is developing a new paradigm for file management, and it seems the front runner for this is some kind of journal, similar to what the OLPC has. Federico Mena-Quintero posted to his blog saying that he believes humans are pretty good at knowing when they’ve done things, and this forms a useful paragidm for interacting with the computer. I don’t know about anyone else, however, but I at least struggle to tell you what day last week I had cabonara for dinner, let alone what month two years ago I worked on that essay which might have been about some aspect of polytheism in the Hebrew Bible.

What Aza Raskin et al would suggest is that there is already a model for interacting quickly and efficiently with large numbers of files and data on the web: text based search. I agree with him too, there’s no way I could work my way through all the billions of pages quickly if we were still stuck using web directories. Not only that, but it’s an interface many people are used to dealing with: my non-technical friends who use the web never type any address out in full except for www.google.com, which they then use to find what ever else it was they were wanting to go to even when it’s something they visit regularly like their email (interestingly, most don’t even realise there’s a quick search bar just to the right that would save them even typing that!).

So, while being able to narrow down a set of search results by date might be useful for pinpointing a hard to locate file alongside other technologies like tagging (a file might belong to two distinct sets) and smart folders,  it seems to me this should not be the default; instead, one of the many existing free software desktop search systems should be used, whether that’s Tracker, Stirigi or Beagle with these other technologies built around search as assistants.

Oh, and search would need to be well integrated into the operating system and not require me to open up a separate application before being able to do anything (as with Tracker now). Maybe something in the panel, maybe something like QuickSilver/Gnome-Do, maybe, crazily, a search bar in the middle of the desktop.

More later, perhaps…

2 Comments

  1. Search? Hmmm. I think they should get the basics right first before worrying about `search’. And that means a stable interactive-peformant desktop that doesn’t get in the way of your work.

    Actually, every ‘desktop search’ feature i’ve tried has just gotten in the way of having a nice interactive desktop that doesn’t interfere with your work. In general it is simply better turning it off and resorting to find and grep for the few times you need it. Rather than having it drag down your system all the time and specifically at times you’d rather it didn’t.

    A journal idea sounds more useful, although it could also become cluttered very quickly. At least you’d have less data to search through. It has privacy and security issues though.

    To be honest, a local machine is very different from the internet, the same search model doesn’t really work. The collected meta-data and group-intelligence you get from link relationships doesn’t exist on the local machine, unless you set it up yourself – which defeats the purpose. And personally I like to ‘forget’ old stuff, or at least, remove it from my immediate conscience – since it is usually of little use.

    Any solution has to sit at a lower level than ‘gnome’ too – many many applications will never be burdened with gnome libraries.

  2. Cant say Im jumping for joy when I read that version 3.0 wont be a ‘real’ break from the 2.x version but that 3.0 makes it seems new.
    There needs to be a major sitdown and figure out how to face the future of desktop and while a GUI change would be nice (my niece calls the Ubuntu on her Dell Mini 9 ‘depressing), I am not referring to superficial changes like Vista 7 has done.

    I want a real x.0 release, not Gnome2.5.

    I cant say I am either impressed by the Mono obsession which has creeped in and have now moved to XCFE full time and KDE for some family who need to be switched.
    And you know what?
    Life goes on…


One Trackback/Pingback

  1. [...] Gnome 3.0/Interfaces in general Across the various planets, there seems to have been quite a lot of discussion about interfaces and how Gnome 3.0 should proceed. I think the most interesting of all these posts has been this one from Alberto Ruiz where he linked to this talk given by Aza Raskin. While I know very little about interface design and human computer interaction, it seems to me that many of his suggestions make a lot of sense and could help to tackle some of the problems people have with their computing experiences. I’m going to share some of my thoughts on this here, take it or leave it, but I at least hope those who are working on Gnome 3.0 interfaces will give some thought to Aza and his father’s work. [...]

Post a Comment

*
*