W3C home > Mailing lists > Public > xproc-dev@w3.org > March 2011

Re: How to add more transports/protocols to XProc?

From: Conal Tuohy <conal.tuohy@versi.edu.au>
Date: Tue, 15 Mar 2011 14:02:19 +1100
Message-ID: <4D7ED6BB.4000103@versi.edu.au>
To: Alex Muir <alex.g.muir@gmail.com>
CC: vojtech.toman@emc.com, xproc-dev@w3.org
On 15/03/11 03:19, Alex Muir wrote:
> I wonder why one wouldn't generally want the steps to execute one 
> after the other unless specified otherwise in someway by the script to 
> for example execute in parallel or to execute and not wait?
Personally I wouldn't want to have to state "these pipelines can execute 
in parallel". As a programmer I want to focus on what are the logical 
dependencies between my pipelines, and any given pipeline processor can 
work out for itself whether or not it can execute them in parallel.

I also don't want to have to physically allocate space on disk to store 
my files, and I don't want to have to do error correction on my network 
connections; it's a similar principle I think. A pipelining language 
should make parallelisation easy for the programmer, and not require you 
to manage it explicitly (in the general case).

The issue arises where you have dependencies that are not captured in 
your code (i.e. step A has to execute before step B, but the dependency 
isn't captured in the XProc script because, as in the example case, it 
only exists in the form of a data file in a file system external to the 
pipeline). In that case, I think Norm's right that the most obvious 
solution is just to "pipe" those steps together, e.g. by "piping" the 
location of the file from the source to the destination.

-- 
Conal Tuohy
eResearch Business Analyst
Victorian eResearch Strategic Initiative
+61-466324297
Received on Tuesday, 15 March 2011 03:02:59 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 15 March 2011 03:03:01 GMT