RE: An unfulfilled requirement maybe?

I've now seen it, and I have some thoughts on it, some of which were vaguely
mentioned before too.

I agree on providing a custom command line interface (as in "a custom set of
rules for executing commands on any command line interpreter"), but if that
interface is to be, it misses a few things I think should be standardized as
well.

Most notably the "command" and "cwd" option. Why not make them URIs? Or
better yet - IRIs. This would also eliminate "fix slashes", which feels odd
anyway. Not to mention it makes it platform independent and consistent with
the rest of the pipeline which uses only URIs and never plain file paths.

Also, I think the environment variables should be dealt with as well. The
exact variables available, if any, should be implementation defined, but the
way by which they are called should be specified. This however calls for yet
another character that needs to be doubled, or escaped in a uniform fashion
(which if established should also apply to quotes and apostrophes).

To solve this, I propose further, windows like simplification, by which any
special character is escaped with "\" and literal "\" are doubled. And the
character to reference environmental variables... how about "$varname$" or
would it be better "%varname%" or just AVT sort of way ({varname})? I'd be
happy with any combo that can allow the variable's value to be embedded in
any literal value, with no space between the string before and the string
after, while still keeping it as stable as the current draft.

How about making the default value of "cwd" the base URI of the option? This
will allow pipeline authors to ship a pipeline with a program (either in the
same folder, or in a known subfolder), allowing the pipeline to work
everywhere the program would, since the path to the program would always be
known in advance.


And I think now is the time to discuss that long delayed naming convention
;-).

"args" to "arguments"?

"cwd" to "current-working-directory"?

"p:exec" to... "p:command-line"? (I still think "p:execute" sounds way too
general, but it would be acceptable too)

-----Original Message-----
From: public-xml-processing-model-comments-request@w3.org
[mailto:public-xml-processing-model-comments-request@w3.org] On Behalf Of
Norman Walsh
Sent: Monday, November 26, 2007 8:53 PM
To: public-xml-processing-model-comments@w3.org
Subject: Re: An unfulfilled requirement maybe?

Vasil,

The next WD will include p:exec which I think resolves this issue.
If you agree, please let us know.

You can close the issue yourself by replying with "[closed]" at the start of
the Subject: header.

                                        Be seeing you,
                                          norm

--
Norman Walsh <ndw@nwalsh.com> | Debugging is 99% complete most of the
http://nwalsh.com/            | time--Fred Brooks, jr.

Received on Tuesday, 27 November 2007 13:41:01 UTC