David, (01)
I completely agree on that point: (02)
> I'm not attempting to be flip or dismissive here, this comment is
> simply for scale, but 1.5M LoC is pretty much rounding error in a
> software portfolio. (03)
The question is not size, but whether the analysis can be automated.
There are many highly automated tools that can analyze programming
languages in order to extract the data structures, control flow,
etc. (04)
Note that a major consulting firm claimed that the task would take
2 years with 40 people to do the job "by hand" -- almost all of
that time would involve people reading documents and relating them
to the previously analyzed and indexed code. (05)
The critical innovation was to use NLP technology to analyze all
the documentation and relate it to the previously analyzed code.
But most software tools rarely use NLP technology. (06)
After the software has been analyzed and indexed, the search
time increases logarithmically. Finding a needle in an indexed
haystack of a trillion lines takes twice as long as finding it
in a million lines. (07)
All you need are two CPUs instead of one (assuming, of course,
that the task of analyzing all the code had been done by
using more conventional software analysis tools). (08)
John (09)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (010)
|