Re: Which libraries in Boost are mature enough to be used in real applications?

From:

David Abrahams <dave@boost-consulting.com>

Newsgroups:

comp.lang.c++.moderated

Date:

11 May 2006 08:35:38 -0400

Message-ID:

<uac9py99z.fsf@boost-consulting.com>

brass@mailvault.com writes:

David Abrahams wrote:

"Le Chaud Lapin" <unoriginal_username@yahoo.com> writes:

In serialization, I noticed that Boost, IMO, came very close, but
then they

Robert Ramey is the author of Boost.Serialization.

started that versioning stuff to compensate for changes in
the declaration of classes...

....and how does the introduction of versioning "create a steady-state
model" of the serialization domain? Do you honestly believe that not
adding versioning would have made a difference in that regard?

--

I think the "versioning" stuff was a mistake also.

You're entitled to that opinion. I don't happen to have a strong one.
I do think the idea that versioning, more than any other arbitrary
part of the abstraction, makes the model "steady state" is nonsense,
and it's that contention I was responding to.

Isn't a utility program that converts data once a better approach
than adding "versioning" code to an application and it being
executed each time you run the application?

I don't know. How do you build and maintain the conversion code?

In the Boost serialization library documentation, Ramey refers to
Microsoft as being a pioneer in "versioning."

OK... not sure how that's related to anything else we're discussing.

Is there a performance cost associated with opting out of the
"versioning" party?

Why would there be? You can just ignore the version number, which
will always be zero.

--
Dave Abrahams
Boost Consulting
www.boost-consulting.com

[ See http://www.gotw.ca/resources/clcm.htm for info about ]
[ comp.lang.c++.moderated. First time posters: Do this! ]