From: David Abrahams (dave_at_[hidden])
Date: 2004-12-01 07:38:11
Jonathan Turkanis wrote:
> "Aaron W. LaFramboise" <aaronrabiddog51_at_[hidden]> wrote in message
>> Vladimir Prus wrote:
>>> How can you rationally decide if a nice colored rocket is better
>>> than b/w logo from Joaquin that is beatifylly minimalistic? It's
>>> simply not possible.
>> I agree that "beauty is in the eye of the beholder."
>> But we may not be looking for beauty. Likely we are looking for a
>> logo that meets a set of requirements, and further performs well
>> versus a variety of useful tests. I do not think objective
>> evaluation of a logo is any more impossible than objective
>> evaluation of libraries to be included with Boost.
> I believe a logo should be simple, elegant, visually distinctive and
> say something very straighforward about Boost, such as that it has to
> do with C++, or has a relationship with the C++ standardization
That may not even be neccessary. Look at Adobe, Microsoft, Apple, UPS, ...
Even FedEx only makes the vaguest semantic connection. I think visual
distinctiveness is really the most important thing. It's more important
that the logo *not* communicate the things we don't intend (e.g.
> The differences between
> (i) Boost expands on the standard library (ii) Boost aims to produce
> libraries which will be standardized (iii) Boost aims to produce
> libraries many but not all of which will be standardized (iv) Boost
> aims to produce libraries of the same quality as the standard library
> are too subtle to reflect in a logo. If a logo requires detailed
> analysis to understand, it's no good. It's like a movie score: it
> makes you feel good, the movie would seem bland without it, but if
> you notice it and have to think about it the composer did a bad job.
> So it seems appropriate to collect a reasonably large number of logo
> proposals with very different designs and eventually vote on which is
> most pleasing.
> The various graphic design requirements that have been suggested are
> a bit too strict, IMO. It's not necessary for the logo to be in
> vector format, as long as high-quality raster versions are available
> in various resolutions. It's also not necessary that the logo should
> be renderable as a 16x16 pixel icon: it should be sufficient that an
> alternate 16x16 logo is provided which has clear stylistic
> connections to the main logo. The requirement that the logo look good
> in black and white seems sound, though.
No real objections, but why?
-- Dave Abrahams Boost Consulting http://www.boost-consulting.com
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk