zhouhui at wam.umd.edu
Sat Feb 5 13:15:23 PST 2005
On Sat, Feb 05, 2005 at 08:32:53PM +0000, Matthew Burgess wrote:
>Hui Zhou wrote:
>>as long as it works and clean, why would "we" care for the standard?
>That sounds so like something a Microsoft employee would write it's
>scary! The whole point of using a standard is precisely because it *is*
What is standard good for? Good for exchange between a wide
applications. To maintain this ability, applications actually
sacrifice for it. For many area, this sacrifice is worth it.
>Why bother using it if we're going to do so in a non-standard
>manner, hence break with everyone's reasonable expectations and
>assumptions of how our tool will act?
I thought everyone's expectations were the usability of alfs, not the
demanding of profile writen in XML with specific complexity and
internal communication protocol.
>I'm beginning to think you're
>only writing these comments to hinder progress. As such, as from now,
>I'm going to no longer waste my time on such meaningless conversation.
If you think so, you should not further hinder progress on intention.
I confess 50% unintentional guilty.
>We've already given you the reasons why we think it remains sensible to
>stick with XML, I don't recall you ever rebutting those arguments,
You feel it sensible is your feeling. How would I rebut that? I don't
really care wheter you stick with XML or not. I only interested in
why you don't see what I see. Anyway, I slowly getting that why now.
>Note that moving to something
>other than XML will require 1) Defining a new format 2) Converting
>existing profiles to that new format 3) Profile authors will have to
>learn the new format 4) Implementing a parser, validator and processor
I think I understand now. It 1) took huge effort of a committee
(pronounce that!) to define XML (write that huge spec), especially
think that they need think for countless usages and do all kinds of
compensations; 2) people are actually *hand* traslating xml documents
3) Profile authors have spent huge effort to learn the old format; 4)
It took DV years to implement libxml2 so he can finally very proudly
claim his library is w3c compliant!
> As such, whatever format you propose better have some *really*
>strong arguments in favour of it to justify that level of work.
The strong argument I have is: with a possible simple approach, all
those difficulties or seemingly huge tasks just remain simple.
More information about the alfs-discuss