I would suggest starting with simple ASCII since classads are simple key
= value pairs. Evaluate the usage patterns and then evaluate whether
XML is necessary.
I have yet to meet the application for which the use of XML made things
clearer.
JT
pierre girard wrote:
> Charles Loomis a écrit :
>
>>> 2) Build a classad parser which does the best it can with the expression
>>> it gets; that parser might be either in the WMS or in the CE.
>>>
>>> 2a) As 2, but insist that users restrict their requirement expressions
>>> to things the parser can deal with.
>>>
>>> 3) Add new things in the JDL which match more directly with the kind of
>>> thing you get in batch systems, e.g.:
>>>
>>> CPULimit = 3600;
>>> MEMlimit = 511;
>>> DiskLimit = 500;
>>> Priority = 3;
>>>
>>> 3) is simpler on the batch system side, but you still have to adapt it
>>> to specific systems, you still have to make it match the glue schema (or
>>> vice versa), and you lose a lot of the potential richness of the classad
>>> expressions.
>>
>>
>>
>> I would instead opt for a hybrid approach of 2) and 3). Allow people
>> to define parameters like in 3) and have whatever processes the final
>> JDL combine those with any explicit requirements to arrive at the full
>> expression. Only those limits given separately would be passed to the
>> local batch system. This makes it easy to parse, loses none of the
>> functionality of the classad, and could be implemented in a finite
>> amount of time. It is also closer to the spirit of the GGF job
>> description.
>
>
>
> I would even opt for a XML implementation approach... It seems to me
> that the problem is basically to rewrite an expression (the JDL source)
> in another one (the LRMS command to be performed). The XML technology
> provides us with a very convenient and a very open way to do this.
> Moreover, there would be no need to implement (yet another) parser and
> we could possibly use an existing XML transformer (i.e. XSL Transformer)
> to finally transform the XML data into the expected LRMS-specific command.
>
> How could it work ?
> the RB could provide the CE with the jdl/classad data expressed as an
> XML document, and the CE could provide something like a XSL stylesheet
> to rewrite the XML data into a local LRMS command expression to be applied.
>
> What would be the implementation effort ?
> 1- At the RB level, to express the classad/jdl (or whatever you want) in
> a XML format.
> 2- At the CE level, to provide an XSL stylesheet depending on the kinf
> of LRMS, which would transform the XML data into a LRMS command.
>
> I don't know the expressivness of the famous classad (is there somewhere
> a classad specification document ?). Then I don't know whether a classad
> expression could or not be expressed as an XML tree... ???
>
> Pierre
>
>
|