xml-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Eric Hodges <harmo...@swbell.net>
Subject Re: parser-next-gen goals, plan, and requirements
Date Wed, 12 Jul 2000 23:57:10 GMT

----- Original Message -----
From: "Arnaud Le Hors" <lehors@us.ibm.com>
To: <general@xml.apache.org>
Sent: Wednesday, July 12, 2000 4:54 PM
Subject: Re: parser-next-gen goals, plan, and requirements


> Eric Hodges wrote:
> >
> > So why is the DOM API so bloated and ugly?  Why doesn't it use Java
> > collections?
>
> Because one of the requirements of the DOM was to be language
> independent. Most companies don't just use one language. They use a
> variety of tools and components that are in different languages. There
> is a lot of benefits for them to have a single API across the board. It
> reduces education costs, support costs, etc...
>
> >  Why are there so many non-obvious steps required just to parse
> > a document?
>
> The DOM doesn't deal with parsing at all for now, so I don't understand
> what this is about.

But there's the problem.  It makes them happy and it annoys me.  There isn't
asingle solution that fits us both.

>
> > I already know why because you told me.
>
> So why do you ask??
>
> > W3C's job was to make several
> > different implementations happy with one API.  The result is an API that
> > doesn't make anyone happy.
>
> That's most often what happens with standards. XML 1.0 itself doesn't
> satisfy everybody for that matter. Does it mean we should through it
> away and start over? I don't think so. Some have tried.

No, but it means some people shouldn't use XML.  And some people shouldn't
use DOM.

>
> The important point is that it's better to have a standard then no
> standard at all.

Or even better, have several interoperable standards tailored to specific
needs.

>
> I know it's hard to accept/understand that. I have an academic
> background; for years I worked as a research engineer where I had the
> luxury of having the time to develop my applications entirely from
> scratch. No constraints at all. I could aim for the best and the
> cleanest of all. Working at the X Consortium and W3C after that, I have
> had to learn the industrial reality. It was quite a chock at first.
> Especially at W3C where I started by working on HTML, by far the most
> controversial place at the time. There I learned that the priorities in
> the industry are very different from what I might dream of. But I also
> learned that there are good reasons for that.
>
> Sorry if this sounds boring. I'm just trying to say that, while I
> understand why people don't like the DOM, they should understand why
> things are the way they are. And why I think throwing it away and
> starting over isn't necessarily the best thing to do.

I understand why it is the way it is.  I don't think it should be thrown
away, but I don't think it's needed in all the places it's used.

>
> > Isn't it too late?  Are they going to throw away the current API?
>
> No, they (we) aren't going to throw away the current API. I don't think
> there is any chance nor any point in doing this (given what I just
> explained). But the WG will certainly consider specific issues and see
> if/how they can be addressed. A read-only DOM is on the list of issues
> to be addressed for DOM Level 3 for instance.

So if someone asks for a DOM for embedded systems, will they consider that?
How about a DOM for Java only?  Or a Smalltalk DOM?

I don't think they can add new standards as fast as I can ask for them.



Mime
View raw message