[xsl] Re: COCOMO project cost metrics for Saxon and XQilla

Subject: [xsl] Re: COCOMO project cost metrics for Saxon and XQilla
From: "Vladimir Nesterovsky" <vladimir@xxxxxxxxxxxxxxxxxxxx>
Date: Wed, 3 Feb 2010 07:36:28 +0200
If to assume some reasoning behind the estimation then it should probably account for specification efforts. In case of Saxon it's a set of W3C specs. If that's the case then it may well happen that algorithm considerably underestimates total investments.

It might be that at the time when COCOMO metrics were designed there was no such an open community that could change estimates so considerably.
Vladimir Nesterovsky

It would easily have taken 50 person years to write Saxon using a
conventional large-team approach - except that the large team would not
actually have delivered the same product.

Solo programming, like solo mountaineering, is probably 10 times more
efficient (but also much more risky) than the conventional large-team


Michael Kay

-----Original Message-----
From: Justin Johansson [mailto:procode@xxxxxxxxxxx]
Sent: 02 February 2010 14:57
To: xsl-list@xxxxxxxxxxxxxxxxxxxxxx
Subject: [xsl] COCOMO project cost metrics for Saxon and XQilla

Hi all,

Having just happened across these COCOMO metrics*** on Ohloh
for Saxon and XQilla XSLT/XQuery/XPath products/projects, I
thought the list might be interested in the "bean count"
estimates assuming a developer salary of $55000 p.a. to write
these projects from scratch.

*** http://en.wikipedia.org/wiki/COCOMO

For Saxon (and presumably this just the Open Source Non-SA codebase)


50 Person Years $ 2,743,467

And for XQilla


35 Person Years $ 1,899,031

Obviously it's anyone's guess if there's more than zero
significant digits in the $ values let alone 7 digits!
Actually I stumbled across the XQilla page first and, given
the friendly URL, took a stab at seeing if Saxon was there too.
And it was!

Any comments, people?

Justin Johansson

Current Thread