[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

SEUL-Leaders: Berlin opinion (fwd)

Haven't read yet, on my way home...  Just thought I'd get it out 'there'.

---------- Forwarded message ----------
Date: Wed, 13 Aug 1997 17:15:33 -0700 (PDT)
From: Sam Revitch <samr@eng2.sequent.com>
To: Erik Walthinsen <omega@eng2.sequent.com>
Subject: Berlin opinion

After reading through their docs, I've formed a few opinions:

Good ones:
* Their server will be multithreaded, and the apps will be using semi-direct 
  graphics command queueing.  This is very good.
* They want to integrate OpenGL functionality.
* They have lots of cool features planned.
* etc.
[In short, they have _all_ the display-end technology straight]

But I have doubts about a few aspects:
* It looks an AWFUL LOT like 'doze.  A short sed(1) script could 
  probably convert 70% of the function names to loser32 equivalents.

* They talk about multi-architecture multi-OS support, but also seem to 
  agree with the idea of using moderate quantities of assembly code for 
  non-hardware-related stuff.  [oops]  I have great doubts in their 
  ability to move this to other platforms with speed.

* It would be pretty ugly to make it work with existing X applications, as
  X programs wait on a socket and the GGI appears to have a black-box
  waiting mechanism (using a semaphore?)  Many X programs do go out and get 
  the file descriptor of the X socket and implement their own waiting 
  mechanism.  Since something like this can't be fixed with a few fancy 
  macros, let alone all the other differences, compatibility could get ugly, 
  strip some features, and add some extra weight.  It's no wonder they say 

As you've mentioned, the lack source-code compatibility with X applications 
will be their fatal flaw unless they address it real soon.  Were the 
native 'doze 95 APIs incompatible with anything previously available, 
'doze 95 wouldn't be quite as popular as it is now.  And Berlin doesn't 
have a $300M advertising campaign behind it.  :)

Perhaps their true aim is to make it easy to port windoze applications to 
Berlin, rather than X applications.  But there are probably fatal flaws 
in that, too.

Summed up, I'm skeptical of their ability to make it popular even with
the Linux community, let alone the rest of the world.  They seem to be 
rewriting the standards as they go.

-Sam Revitch

Simple End User Linux Leader Mailing list