On May 09, 2012, at 01:57 PM, Suryajith Chillara <[email protected]> wrote:

1. Choice of language for the web framework : php or Python.
[snip]
The front-end however shall use php as it has to be written as a mediawiki plugin(The exact details could be discussed later as per the roadmap[6] mentioned) and it shall be decoupled from the web API service.

I think that detail doesn't really matter for the backend.  You could implement a fantastic or terrible system in either.  It sounds like you're more familiar with and interested in using Python and I don't see a problem with that.

I don't think there's value in implementing your own CMS on the front-end so therein is where there's a desire to have the site integrate with some existing framework (whether mediawiki, drupal, or something else).  Bridging that front-end through something like Apache Thrift, however, sounds like a fantastic idea.  I'd probably be more of a fan of that over Bottle just since that would mean we could make interface clients in other languages (e.g., C++) more easily.

As for your diagram, it would be nice if all processing could go through the queue instead of allowing the web API to circumvent it.  That will support improved scalability should the need ever arise, not lock up the webserver handling requests (processing can be fully asynchronous), and simplify the architecture a little bit.  If someone encounters a bug a couple years from now in the web API, we know it's not related to processing and should be able to isolate problems more easily. 

2. Data to be captured :

The data that benchmark script currently captures includes machine description, version of brlcad installed, time stamps, run time, default values needed to run the benchmark, system info, processor info along with the performance metrics of the run. Along with that Sean wanted the tool to capture the compilation flags, the version of glibc used etc. I shall send in a mail *soon* discussing the db design to improve the db schema mentioned in the proposal[1]. I would also want to know what important information that you'd want me to capture.
 
Basically "everything available" properly normalized and indexed, including the actual log file as-is.  It should be just sitting in a directory somewhere in case we ever want to automatically "repopulate" from scratch or manually inspect.

Cheers!
Sean

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
BRL-CAD Developer mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/brlcad-devel

Reply via email to