[capidl] Re: [E-Lang] Concerning XML docs
Jonathan S. Shapiro
Tue, 18 Sep 2001 20:04:13 -0400
> Can someone explain to me why we are having anything to do with the whole
> XML cargo cult in the first place?
> Is there any actual technical benefit from touching this tarbaby or is it
> done for marketing/positioning purposes?
I'll take a cut.
Regardless of the merits of XML, it's clearly useful to have a means of
content tagging that is generally agreed on. Separately, given a usable but
potentially imperfect solution that is widely used and supported, it is
generally preferable to use that solution than to reinvent unless there is
*compelling* reason to change.
XML is certainly not perfect, and we can all imagine better tagging schemes,
but there is a large class of jobs that XML gets done, and it is definitely
At the risk of fanning various peoples' flames here, I have observed in the
E group where XML is concerned a tendency toward well-motivated but ruthless
purism, as expressed (for example) in the rejection of XML in favor of
MinML. Some of the objections to XML are completely valid but not really
important in practice. Some are simply silly. A few issues may actually
matter, but I don't understand whether the truly impact us in practice.
In the process of this quest for purity, I think that we have lost sight of
the importance of wide acceptance. MinML is a class of valid XML DTDs, but
it violates the *expectations* of the XML user base. Telling them "but its
XML" may be factually accurate, but it isn't relevant. When we decide to use
MinML ourselves for our own DTDs I have no problem. When we restrict the
inputs that we will accept from users and third-party tools to MinML, we are
engaging in balkanization whose cost considerably outweighs (IMHO) any
I understand, and probably agree with, arguments of the form "we have too
much to do right now to look at this issue." It's the willingness to ignore
the user base that bugs me.
I probably shouldn't have posted this, but I wanted to re-inject the idea
that this decision is just like the E bracketing syntax decision. The
question should be not "what is right", but "what does the user expect and
can we live with it?"