Boost logo

Boost :

Subject: Re: [boost] [math] calculate derivative automatically.
From: Ross Levine (ross.levine_at_[hidden])
Date: 2009-03-11 11:11:57

I am going to have to agree with Hervé. If I want a derivative, I can
use maple: diff(function, variable). I just don't see templated
differentiation as all that useful. As an engineer, if I'm using C++
for a task that needs function evaluation, all derivatives will be
precomputed. Also, you have to take care of things like pow(), exp(),
log() and log10(), sqrt(), repeated fractions, etc. It would be
possible, but ultimately, it's going to take forever to compile and it
probably just isn't going to be that useful to most people.

On Wed, Mar 11, 2009 at 7:58 AM, Hervé Brönnimann
<hervebronnimann_at_[hidden]> wrote:
> No, your messages are coming through fine (4, 6, 8, and 9 of March).
> I think the lack of response might be indicating of the lack of interest,
> but people are too polite (or too busy) to tell you in such blunt terms. I
> almost hesitate to do so myself, because I wouldn't want you to be
> discouraged or feel like you've been dismissed out of hand. I've always
> found the people on boost to be the most courteous, and wouldn't want to
> depart from this honorable tradition. I've also found that there are a lot
> of people out there who may not share my opinions or interests, and that's
> fine too. But I see you're starting to invest valuable time and might not
> appreciate not being told in the long term.
> Speaking for myself (and myself only) from now on, I didn't bother replying
> because it seems to me that you are simply misguided. As someone said, it
> is technically feasible. Almost anything is. But that doesn't mean it
> should be done (esp. bearing in mind the cost of developing a good solution
> to your problem). I appreciate that the engine may work for you but it's a
> long way from being of general interest, and even if it were, I doubt it
> would belong to boost. It's very likely that the simplification step will
> be much more complicated than you think. Already, even commercial products
> (I played a bit with Maple) find this to be a very hard problem. And the
> techniques they use, including unlimited precision integers, numerical
> methods, Grobner bases, etc. are all *way* out of scope for something like
> boost. Anything simpler-minded is likely to answer a much simpler problem
> for which the best solution is simply to compute the derivative by hands (or
> non-C++ means) and plug it in.
> If I really want to invest more time into this problem, would strongly
> recommend you try contacting other, more specialized, mailing lists, like
> yacas if it is still in active development.
> Again, those are my opinions only, for what it's worth.
> Respectfully,
> --
> Hervé Brönnimann
> hervebronnimann_at_[hidden]
> On Mar 11, 2009, at 6:41 AM, Dongfei Yin wrote:
>> BTW: I am wondering whether I do something wrong, there is no responding
>> about this subject for a long time.
>> Dongfei.
>> _______________________________________________
>> Unsubscribe & other changes:
> _______________________________________________
> Unsubscribe & other changes:

Boost list run by bdawes at, gregod at, cpdaniel at, john at