W3C home > Mailing lists > Public > xsl-editors@w3.org > January to March 2000

Re: Question on Transformer for Building OracleTransformer

From: Scott Boag/CAM/Lotus <Scott_Boag@lotus.com>
Date: Thu, 13 Jan 2000 20:13:22 -0500
To: cocoon-dev@xml.apache.org
Cc: xsl-editors@w3.org
Message-ID: <OF13FB8431.A21AD7ED-ON85256866.0004302E@lotus.com>

Stefano Mazzocchi wrote:
>> > The very first line of the spec says that XSLT is a file to transform
>> > XML documents into XML documents.
>>
>> Where?  I don't see that in my copy of the XSLT recommendation.
>
> Well, very first line of the Abstract...

Ah, so it does (I was not looking at the abstract).  I'm very surprised
that it is there.  I consider this a bug (notice my cc).

> On the other hand, how can XSLT presume to
> identify all the possible results of the XSLT transformations? in fact
> (XML, HTML and TEXT) are not partitioning the result space since both
> XML and HTML are TEXT.

Huh?  It doesn't.  For one, notice that the method attribute is a qname.
For another, as I said, xsl:output is there to effect the serialization of
the result tree, which is really an intermediary result or a type of
diagnostic, at least for anything that is doing formatting.  xsl:output
shouldn't ever effect the semantics of the result tree.

> What to know my opinion? <?xml-stylesheet?> should be an HTTP header and
> <xsl:output> should not exist.

Well, the WG's job is to listen and respond to what people need, and people
needed and wanted xsl:output.  Before we had it, these same features were
being done by the processor, and the processor had no idea of the intention
of the stylesheet author.

As for xsl-stylesheet, I agree, and I think most people involved would too.
Except, again, it's a hint and useful for some things.  Remember this stuff
is for generic use, and cocoon isn't the only consumer of these designs.  I
think a lot of what people want to do with XML, and XSLT is ugly and
unpure, but if I take the attitude that everybody but me and the people
that agree with me are idiots, I don't think I could do this work at all.
All of XML and XSLT is one big compromise.

> Right, and this is exactly what we are going to do: ignore xsl:output.

Absolutely.  It is an instruction to the serializer, and the serializer
only.

-scott
Received on Thursday, 13 January 2000 20:14:24 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:59:50 GMT