Re: [ublas] Aliasing for += and -=

From: christopher diggins (cdiggins_at_[hidden])
Date: 2005-06-13 08:43:53

----- Original Message ----- 
From: "Kresimir Fresl" <fresl_at_[hidden]>
To: "ublas mailing list" <ublas_at_[hidden]>
Sent: Monday, June 13, 2005 1:51 AM
Subject: Re: [ublas] Aliasing for += and -=

>
> christopher diggins wrote:
>
>> [...] This is trivial to
>> fix. I find it unacceptable that the library can not do such a simple 
>> task
>> correctly.
>
> Would you be kind to browse through old posts to ublas-dev mailing list
> (http://groups.yahoo.com/group/ublas-dev/) before making such dismissive
> statements?

There are over over 2500 messages. I don't have the luxury to "browse" 
through these messages for no specific purpose. If you want to point me to 
something specific I would be happy to read it.

I did however do my homework which I think earns me the right to make a 
dismissive statement. I read the documentation (what was available, it is 
still broken), I read the effective ublas wiki entry, I read gunter's guide, 
I asked questions first, and I did a detailed comparison of performance of 
ublas with two separate matrix implementations, along with std::valarray.

My point is simple: any matrix/vector library (especially as part of boost) 
should at the very least be able to perform better than a naive 
implementation, or a standard library implementation (std::valarray), in its 
most intutitive form (for instance v += u) without requiring the user to 
read reams of documentation, ask questions on newsgroups and mailing lists 
or have to use extra functions.

Christopher Diggins
http://www.cdiggins.com