ant-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Steve Loughran" <>
Subject Re: Ant: The Definitive Guide (Orielly)
Date Tue, 28 May 2002 19:59:36 GMT

----- Original Message -----
From: "Chris Bailey" <>
To: "Ant Users List" <>
Sent: Tuesday, May 28, 2002 11:08 AM
Subject: RE: Ant: The Definitive Guide (Orielly)

> Ok, I'll be looking forward to this book :)  I agree on printing the Ant
> reference in the Hightower/Lesiecki book.  That's just a waste of paper,
> is instantly out of date.  Too many books do this, and should realize in
> this day and age, unless the book itself is a desktop reference (and only
> that), that it's pointless (and even then it's pretty debateable, unless
> it's like say the Java reference books that are more than just JavaDoc,
> give nice examples).

I am never a fan of regurtitated documents. But it is useful to have a hint
of reference so that you can turn to them while reading the rest of the
book, or are in that off-line-no-laptop state that I have been in for the
last two weeks, since the untimely death of mine.

Ant: tdg must have expanded or reworked the docs, not based on sight but
simply on scheduling and differentiation reasons: they would have been out
early with a product that just matched the on-line docs if they hadn't. So
it may well be a good alternative to the 1.4.1 docs, explaining stuff better
and with a hundred and twenty plus intro-to-ant section at the beginning. We
shall have to see.

One thing I am going to be curious about is whether they went back to the
source, versus drove off the docs. If there is one flaw of any open source
project it is that the docs are treated as secondary to the source, since
there is always the source there if you need it. In ant's case, obscure
attributes often get left off, dependency checking behaviour is never
stated, and other sublties left out. As an example, it was only when I
looked at the <echo> source, that I discovered it has had a level parameter
since before ant1.3, but because it was undocumented it was unused. Thinking
about it, echo probably makes a good quick test: if it has the level
parameter then either they did look at the source, or they were tracking
ant1.5 changes enough to notice when I did write up the attribute.


To unsubscribe, e-mail:   <>
For additional commands, e-mail: <>

View raw message