[boost] Code structure analysis

From: David Abrahams (dave_at_[hidden])
Date: 2007-04-26 19:13:47

Reposting for Paul with a better subject line


attached mail follows:

Excuse the super long delay in posting this, I did try a couple of weeks 
back but failed to register first.

David Abrahams wrote:
> on Mon Apr 02 2007, Roland Schwarz <roland.schwarz-AT-chello.at> wrote:
>
>> David Abrahams wrote:
>>> No, boost is not even one library much less a monolithic one.  Boost
>>> is a library collection.  I've seen the dependencies among the Boost
>>> libraries graphed, and unlike some other development frameworks,
>>> which just show up as a big blob, Boost is neatly partitioned and the
>>> dependencies tend to be a unidirectional DAG.
>> Really? I very much would like to see this graph. 
>
> It was in a preview of code structure analysis software from this
> company: http://headwaysoft.com/index.php.  They were interested in
> making an analysis of Boost publicly (I think) available online.  I've
> Bcc'd my friend there; if he wants to pipe up, he will (Paul, you have
> to register if you want to post on this mailing list).

Just back from vacation so excuse the slow response.

>> If this is available somewhere I think it would be nice to show it
>> somewhere on the web!  

Dave was referring to a product we used to sell called Headway Review, 
which had reasonable GCC C++ support and poor VS C++ support. Headway 
Review has since been put out to pasture, and been replaced by the 
leaner, meaner, cleaner Structure101.

Support for C++ for Structure101 is long overdue, but hopefully will 
see light of day before the end of the year. More anon.

Given the lessons we learnt from Headway Review in general and Headway 
Review for C++ in particular, when Structure101 for C++ does finally 
see the light of day, I am confident the folks in the Boost community 
will be interested. ;-)

> Agreed, for real.
>
>> Besides being a nice eye-candy it would
>> support those in need for these kind of arguments.
>
> Ditto.
>
>> I would have volunteered in providing this graph, but unfortunately 
>> I must have been doing something completely wrong when I was trying 
>> to generate such a graph on my own. It reminded me very much of the 
>> wool ball my cat likes to play with than a well structured graph. :-(

If you really want to do this now, I'd be happy to provide an 
unsupported copy of Review so you can get some sense of where you stand.
Contact me directly if you do.

> Well, at the time, Headway's software really did show a remarkable
> difference between Boost and ACE in that respect.

Thats for sure! ;-)

_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost


-- 
Dave Abrahams
Boost Consulting
http://www.boost-consulting.com
Don't Miss BoostCon 2007! ==> http://www.boostcon.com