Re: Transformation + FOs makes abuse easy

Subject: Re: Transformation + FOs makes abuse easy
From: "John E. Simpson" <simpson@xxxxxxxxxxx>
Date: Wed, 28 Apr 1999 14:36:49 -0400
At 12:54 PM 4/28/1999 -0400, Simon St.Laurent wrote:

	[various points]

First, Simon, *thank you* for changing the subject of this thread... this
one is much less, shall we say, prone to inflammation. :)

Since the flurry of responses to the original "FOs considered harmful"
post, I've pretty much shut up -- trying to see how people of such sound
mind on both sides of the argument could have such different
interpretations of "what is." I think Simon's initial posting under this
new thread goes a long way to making clear (or -- at least to me --
clearer) the point of the criticism of the XSL FOs.

Nevertheless I still don't get it. The argument, at root, seems to rely on
an assertion that the W3C is about to bless the emission (in the sense of
"publication," whether of document- or data-based XML) of pure FOs. (Anyone
who wants to can do whatever they want to *without* the W3C's blessing, of
course -- including generating XSL FOs, PDF, or cuneiform.)

But that assertion is false, because the W3C does *not* bless the emission
of pure FOs and the discarding of semantically rich source documents (which
is required in order to make the case for "FOs considered harmful"). The
spec says that you can use XSLT to produce FOs, or you can use XSLT to
produce something else. The latter case is outside the scope of the
discussion. In the former case, to the extent you've "lost" semantics,
you've "lost" them in the client -- assuming XSL-aware clients. If you
choose to produce static FO-only documents for distribution from a server,
then -- duh! -- no one will be able to search your site in any meaningful
way; your site will be less accessible; and all the rest. Like the saying
goes, you can't legislate morality -- and I daresay you can't legislate
stupidity either.

>The growing separation between the
>transformation tools (which have uses beyond FOs) and the formatting
>vocabulary (which is now another XML application) has created a very real
>possibility that organizations will choose to send their XSL-processed
>information to browsers using the FO vocabulary or presentation-oriented
>HTML.  

The separation AFAICT is such only to the extent that XSLT can now
officially be used on its own. Again AFAICT, FOs can be used only in the
context (as the output) of a transformation.

>...the 'meaningful Web' project that was the driving force (at
>least in public) for the creation of XML is at risk.  Server-side
>transformation from semantically rich private vocabularies to
>presentation-oriented public vocabularies may leave the Web exactly where
>it was before - interesting to read, but not very useful. 

And the dopes who've produced such sites will get exactly what they
deserved. Useless pages will get no visitors, and their proprietors won't
be able to exchange data in any meaningful way with anyone else. (You can
also write a DTD using nothing but "ANY" content models. Pretty stupid for
doing almost anything real, but where's the "'ANY' content models make
abuse easy" firestorm?)

When I was a kid, watching as my Dad puttered around under the hood of the
Dodge, I once "helpfully" pointed out that if he didn't have a Phillips
screwdriver handy, he could use a plain-old slotted screwdriver instead. Of
course (as it didn't take me long to learn), you can successfully do so
only in a limited set of circumstances -- when the screw's not too tight to
begin with, and/or when you don't care about stripping the edges of the
little Phillips "+" and hence rendering it useless thereafter. The range of
circumstances in which you can *un*successfully do so, yes, is much greater. 

But geez, I'd hate for the SAE to outlaw slotted screwdrivers.

=============================================================
John E. Simpson          | It's no disgrace t'be poor, 
simpson@xxxxxxxxxxx      | but it might as well be.
                         |            -- "Kin" Hubbard


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list


Current Thread