Endtag attributes

[Anders W. Tell]

Rick Jelliffe wrote:

> > From:   Anders W. Tell
>
> > Im especially interested in arguments for having endtag arguments.
>
> One good reason pro is because when you are doing text processing using a
> streaming tool (i.e. one that does not load the document in memory)
> sometimes you wish to write out attributes with values collected from
> processing sub elements.
>
> Consider an application which reads a table and adds an attribute with the
> column and row count.
>
> This is a kind of foward-reference problem. In streaming text processing
> tools you:
>
> * have two passes, or
> * write out the values of all the attributes in an external entity file (so
> that the XML parser resolves the references on next document load), or
> * have a built-in reference resolving phase such as OmniMakr uses.
>
> Until recently (and for all I know, maybe still) most text processing of
> non-HTML SGML/XML marked-up files was done using streaming tools (i.e. with
> event programming).

Is possible to save parts of the document im memory before writing it to the
next stream in a chain. This may not be desirable in cases when documents are
large .


> The trouble with having end-tag attributes for this kind of use is that even
> though it would be easier to generate such documents, an end-tag attribute
> solution does not buy you anything, because if you are using streaming tools
> you will still need the "collected value" attributes when you come to the
> start-tag, in order to set up processing of the rest of the element. You
> still need some step to move the end-tag attribute to the start-tag.

Not all attributes , which are calculated after processing a fragment, are usable

before the all other attributes and children have been processed. An example
of this type of attributes is "Signature" which can may be read when the
start-tag
token arrives but the validation of the signature cannot be perfomed until the
endtag arrives.

However I agree with you that most attributes dont add anything when added in the
end-tag.

> And, as has been mentioned, maybe people who are used to tag-based text
> processing rather than element (i.e. range) based processing will not be
> encouraged to alter their thinking, if end-tag attributes were allowed.

True,

>
> Perhaps the one thing that end-tag attributes might be useful for is
> checksums, though.

And its alter ego Signatures :)


/Anders
/_/_/_/_/_/_/_/_/_/_/_/_/_/_/
/  Financial Toolsmiths AB  /
/  Anders W. Tell           /
/_/_/_/_/_/_/_/_/_/_/_/_/_/_/



xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)