X11 bashing

[Ashmeet S Sidana]

In article<558 at appserv.Eng.Sun.COM> lm at slovax.Eng.Sun.COM (Larry McVoy) writes:

>jeenglis at alcor.usc.edu (Joe English) writes:
>> peter at ficc.ferranti.com (Peter da Silva) writes:
>> >The problem is that [X's designers] were factoring the problem 
>> >apart along the wrong
>> >lines. They implemented basic drawing primitives and assumed that was good
>> >enough. What they needed to be implementing was visual objects: buttons,
>> >text panes, windows, etc.
>> 
>> I think this is one of the things X definitely does right.
>> It allows for much greater flexibility in UI style and policy.
>> X is still used extensively for UI research, so this flexibility
>> is important.
>
>I think that this is a trap, a typical Computer "Science" sort of pitfall.
>All your college professors will tell you about separation of policy and
>mechanism like it is some sort of manna from heaven.  In the area of
>user interface this is, in my opinion, the worst possible thing that 
>could be done.  It is worse than having a bad, but consistent, user
>interface.

But herein lies the problem. X was NOT designed to provide a
user-interface.  One of the goals of the original design WAS
separation of policy and mechanism. So saying "X did it wrong" is
incorrect. What they set out to do they achieved.

However, later, X started being used for something else - A commercial
standard to build distributed application interfaces with. Why?
Because of interest from the big computer vendors (HP, DEC etc.)  in
the mid-80's. For various reasons NeWS wasn't *commercially*
acceptable, they needed something to counter it straight away and so
they pushed X.

Now, In the commercial world it makes sense to have consistent user
interfaces etc. so the direction of X changed. I have been programming
in X for a long time and I can attest to its change. Release 10 didn't
even have a toolkit! Now you have too many!

Will X ever be as good as something designed from the ground up now.
No! But that is the penalty of retro-fitting something and the
march of technology.

>Think carefully before you flame me - think hard about the Mac.  The reason 
>that *users* like the Mac is due, in part, to the consistent look and feel
>of the user interface.  You may not like it, but you remember how it works.

Good point.

>
>X blew it by handing out all that mechanism to developers.  It would have
>been much better if they took a little longer and came up w/ the same
>set of functionality that the Mac (even the early Mac) had.  Then all the
>apps would look the same, work the same.  The toolkits were only a weak
>attempt.

But thats what they were trying to do! Toolkits for building interfaces
were an afterthought due to "other" reasons.

Oh, I've have completed a loop here. Now the computer can handle the
subsequent iterations :-)

---Ashmeet Sidana
   sidana at cs.stanford.edu
   sidana at hpcc01.hp.com

PS: My last comment is an attempted pun on Alan Turing's (THE turing)
    arguments on computable methods. Flames on that to /dev/null. 
    Just finished reading his biography "Alan Turing: The Enigma".
    A wonderful book!