[boost] Re: Variant Library Review

From: Eric Friedman (ebf_at_[hidden])
Date: 2003-03-06 18:06:37

Andrei Alexandrescu wrote:
> By the way, any chance to tweak the acknowledgments a bit. They mention
> "Generic<Programming>: Discriminated Unions" Part 1, Part 2, Part 3. C/C++
> Users Journal. 2002", but the text doesn't mention the earlier
publication,
> which is
> http://oonumerics.org/tmpw01/alexandrescu.pdf. I think that's where
various
> kinds of visitation in the context of C++ discriminated unions were first
> considered a core feature, together with a number of more obvious things
> (that made it or not), such as using typelists, computing alignment, and
> dispatching via pseudo-vtables.

Can do. We had excluded this since you reference it in your later articles.
But I suppose it makes sense to acknowledge the history.

> Also, the ack text doesn't say whether the work above influenced at all
the
> library design or is just previous work developed unbeknownst to the
> authors - the Generic<Programming> reference is specified in the
> bibliography as [Ale02] without being referred to at all.

Sorry this is a typo that was carried over from an old version of the docs.
Where the acknowledgments read:

  "Andrei Alexandrescu's work in [Ale01] inspired the library's design."

It should instead reference [Ale02]. The problem was that I mistyped the
publication year for the Generic<Programming> articles in an earlier version
of the docs.

So I will fix this. As well, I will update the docs to reflect the influence
of the oonumerics paper on the design of the library.

> If the authors were aware of the previous work and used it as a source of
> inspiration, it's nice to give credit where it is due. It costs nothing
and
> it is considerate. Just recently I failed to give due credit and boy it
> didn't feel good.

As I said above, this was the result of a typo and certainly not intended to
acknowledge your prior work. (Indeed, the sandbox CVS history shows this.)

Sorry for any misunderstandings.

Thanks,
Eric