Older blog entries for haruspex (starting at number 130)

25 Oct 2004 (updated 25 Oct 2004 at 19:13 UTC) »
mathrick, we are obviously talking about different things, and therefore completely at crosspurposes. I am talking about user interface (which by definition has nothing to do with specific platforms [sigh]); you seem to be talking about a set of incompatible rulebooks. It's those incompatible (and totally unnecessary, btw) APIs, aka rulebooks, that contribute to the fragmentation that I am complaining about. Your argument is like saying there is no such thing as a cross-platform algorithm because if you write it in BASIC, a C compiler gives errors!

Google is a good example of the antidote, and gives a glimmer of hope because it uses the kind of abstracted description (HTML) that I am talking about and does, consequently, run nearly anywhere useful. Web applications are ubiquitous because they literally can be written once and deployed anywhere. And that is the holy grail. But HTML isn't the final answer, for obvious reasons. The original hypothesis is that there is not yet a final answer for a platform-independent UI. And I believe the answer probably lies in a description (like HTML) not an API (like all the rest). One reason is because it decouples UI from implementation language. Is that so hard to understand?

Swing is still yet another API, and, I therefore claim, a dead end, like the others (some listed below roughly in order of least useful to most useful).

  • Win32, tied to Windows
  • Cocoa/NS, tied to OS X (and ObjC into the bargain)
  • Carbon, tied to MacOS
  • Xlib, tied to X11
  • Swing/AWT, tied to Java
  • GTK/Qt/wxWindows, their own platforms
  • HTML, tied to the web browser (not the web)
  • 24 Oct 2004 (updated 24 Oct 2004 at 23:31 UTC) »
    There is *no* pure-technical mechanism (framework/API) that can cover the differences between platforms
    That's exactly what I said. But it is easy to imagine one*.

    (Of course, there are frameworks that are platform neutral (wxWindows, Qt, GTK, Tk) although I have hardly used them, and clearly one of those has not predominated as an obvious panacea yet. I suspect a non-API "UI description", perhaps OO, may fit the bill eventually.)

    *Because, as I mentioned, the UI - as seen by users - is by definition not concerned with platform quirks. Nor is program logic in general. And the ideal is not to waste people like raph's time with the dead-end coding of platform specific drudgework.

    (update) mathrick:

    TAINTACPUI - There Ain't No Such Thing As Cross-Platform UI.
    That's exactly what I said. But it's easy to imagine one.

    I hereby bow out of the discussion, I'm tired of being misread. [If you want to get closer to my meaning, forget about "APIs" (which are extremely boring) and think about "descriptions". The description is semantically platform neutral - again, by definition (sigh) - as is, for instance, the description of an algorithm. A platform-indifferent description of a UI is as equally possible as, for instance, platform-independent code, which we all write daily.]

    (update 2) salmoni, thanks for providing an example I was casting about for - HTML is a description, not an API... Smug naysayers take note.

    23 Oct 2004 (updated 23 Oct 2004 at 21:50 UTC) »
    Another platform, another GUI
    While currently you would have to rewrite the UI for each platform to get the nicest result (as several people have pointed out), this is, in principle, unnecessary exertion: The user interface is (by definition) platform independent, and it is very easy to imagine a high level description of the user interface that isn't tied to any particular framework (not necessarily an API). Since such descriptions aren't mainstream (yet?), the problem currently becomes finding a framework that runs on all target platforms. At least you can then work to a single API. If a suitable one can't be found, or the compromises are unacceptable... yes, the effort must be duplicated... D'oh!
    21 Oct 2004 (updated 21 Oct 2004 at 21:35 UTC) »
    REP - someone who wants to improve their own lot (at expense of the world), or maintain their comfortable, unsustainable status quo;

    DEM - someone who wants to improve the world.

    Seems to apply equally to FOSS workers vs capitalistas (M$, SCO).

    Sick of stupid, bloodthirsty, self-interested, warmongering fools.

    Tired of seeing comments such as "liberals are Utopians", "liberals are living in the past". Liberal is a word now as meaningless as "terrorist"; both are used to mean shades of "someone who disagrees with [my conservative views]".

    I am not afraid to inhabit the world and history both. The REPs can have their cesspit Post History, New World. They created the ugliness. They can live in it, or let the humanists fix it (as always).

    And this is disgusting (State Dept roundly and hypocritically insults Castro and all Cubans). From what I hear of the two countries, for the non-rich, life in Cuba is often more comfortable than life in the US. But the US govt would never let the facts get in the way of a good opportunity for some astonishingly brazen propaganda, even by their standards. I suppose the imbeciles in public office think ignorance will always be the controlling force in America; I hope, for humanity's sake, they are not correct. Please vote.

    nymia, love your reflection on History! LOL.
    18 Oct 2004 (updated 18 Oct 2004 at 15:40 UTC) »
    nutella, scary. I knew the world was going to pay for Beavis & Butthead eventually.

    raph, I too was underwhelmed by the results of the paper. I don't share the authors' confidence in their approach but I see how seductive it might be to a non-type-designer... the same intellectual trap that mars METAFONT results, as you say: the conceit that letters can be dissected into mathematical objects at a high level. The research seems to show that this is not the case. I did think the paper had some interesting aspects but these were tangential (haha) to its main premise.

    Re: UI frameworks, yes, it's a tragedy that we have to wrestle with several mutually incompatible UI models. Sigh. Guess each player doesn't really want cross platform software (excepting the open source toolkits of course).

    15 Oct 2004 (updated 15 Oct 2004 at 02:28 UTC) »
    raph: Nick Summers sent me a link to this paper, Parameterizable Fonts Based on Shape Components - have you seen it?

    Also these results.

    13 Oct 2004 (updated 14 Oct 2004 at 04:19 UTC) »
    raph, "native X server"? What about a Quartz version? (and/or libart/X11)

    Gah. Does anyone else here hate D-Link? I have a D-Link DWL-122 USB compact wireless thing in an iMac that gives grief to two other devices on my Ad-Hoc wireless net:

    1. a P4 running WinXP with a D-Link DWL-G520B PCI card can't access Samba shares on the iMac or surf reliably (the iMac is gateway); can't ftp either.

    2. a P3 laptop running Linux can't do certain transfers (http, nfs) with the iMac over wireless. Funnily enough, ssh and ping work perfectly.

    Hmm...wonder if it's MTU/MRU related.

    Gah. I think the next thing is to swap the D-Link USB for a Linksys USB.

    (update) Ankh, NAT is not involved. MTU changes do not help... I think I am dealing with a faulty D-Link USB wireless adapter. Will probably buy a Linksys wireless router tomorrow (only $60 approx!) and solve it that way...

    (I wish Advogato wouldn't keep spawning <p>'s every time I edit a post)

    Found among M$ blogs.

    1) Andy Rich says, "I love when customers find bugs!"

    He must be a happy man then. Maybe one day he'll realise it's even better when customers find and fix bugs.

    2) Brian Johnson's blog is titled BufferOverrun. Bad omen, or what? Sure enough, the top post is a Security Bulletin - the GDI+ vulnerability.

    9 Oct 2004 (updated 9 Oct 2004 at 14:49 UTC) »
    National tragedy in Australia

    Australian election is a disaster. That's it. I'm leaving the country. I can't believe they re-elected that pathetic drooling liar for another term.

    121 older entries...

    New Advogato Features

    New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

    Keep up with the latest Advogato features by reading the Advogato status blog.

    If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!