[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: The relentless march of abstraction

• From: Dave Winer <dave@u...>
• To: John Cowan <cowan@m...>
• Date: Mon, 26 Feb 2001 20:47:00 -0800

> The rule that XML parsers just give up if there is a violation of
> well-formedness.  HTML parsers never give up, and just keep on
> truckin'.

Got it. This is one of my favorite things about XML. Maybe "forgiving" was
the wrong word, perhaps "kind" would be more what I was thinking about.

About the "potted" business -- I truly am confused by much of what goes on
here, so it wasn't potted, I thought I finally got infosets, and was
excited. I didn't get them? They seem like a DOM but for data, a
programmatic way to walk a tagged-text-derived tree. I thought everyone did
this, we do, it's an internal representation of XML text that scripts can
walk almost as if they were part of our object database. We had to bend over
a little to make it work, it wasn't a perfect one-to-one matchup with our
hierarchy, but it was pretty close. I can see why this is very hard to do in
a relational db, and I would never want to program arbitrary XML content in
a relational model. I wonder how much of what's going on here is an attempt
to make an arbitrary hierarchy map onto tables.

Anyway, I've been a hierarchic guy ever since I started programming so many
years ago. I'm glad my bank uses a relational db. I wouldn't want to make
plane reservations using an object database. So I get why it's important.

Dave

• Follow-Ups:
  • Re: The relentless march of abstraction
    • From: John Cowan <jcowan@r...>

• Prev by Date: Re: Why not reinvent the wheel?
• Next by Date: Re: Why not reinvent the wheel?
• Previous by thread: Re: The relentless march of abstraction
• Next by thread: Re: The relentless march of abstraction
• Index(es):
  • Date
  • Thread