Ruven Brooks:
I also wonder how big the spec would be if cast in a formal notation - an order of magnitude larger?

Well, it would be a pretty lousy notation that didn't allow for abstraction and meta-descriptions to cut things down a bit -- unless there are no two parts of the product sufficiently similar to benefit from this, in which case good luck writing the user documentation.

The size of this spec and of other specs raises an interesting issue.  The underlying 
model, going back to the
early 70's, is that people writing the code will be given a specification that tells 
them exactly what code to
write and, therefore, need to understand little or nothing of the application for 
which they are writing the code.

Where I work, we call these people 'compilers'.


Of course, in fairness to Microsoft, one needs to ask is this really one specification 
or is it a collection of
specifications that can be understood in isolation - is it one novel or is it more 
like the Harry Potter series?

The article actually states that there are about 4,000 individual specs, each of between 30 and 50 pages, which, taken together, specify the product. I take this to be a generalisation of the real position, since at least some features in Office are clearly much more complex than others.

Despite this weight of spec, the article also states that Microsoft still gives its 
developers
a lot of leeway for improvisation twixt spec and code, because "they're smart people."
--
Frank Wales [EMAIL PROTECTED]


---------------------------------------------------------------------- PPIG Discuss List ([EMAIL PROTECTED]) Discuss admin: http://limitlessmail.net/mailman/listinfo/discuss Announce admin: http://limitlessmail.net/mailman/listinfo/announce PPIG Discuss archive: http://www.mail-archive.com/discuss%40ppig.org/

Reply via email to