Re: [boost] Why use Boost?

From: David Abrahams (dave_at_[hidden])
Date: 2006-02-03 07:51:51

Sebastian Redl <sebastian.redl_at_[hidden]> writes:

> Paul Giaccone wrote:
>
>>The issue is this: the current documents say "Library X does Y", or, in 
>>functional notation, y = foo(x).  If you know x, you can find y.  
>>Unfortunately, the new user knows y, and so needs x = bar(y), where 
>>bar() is the inverse of foo().  (What's more, the user doesn't even know 
>>that they need x = bar(y) because they don't know that Boost deals with 
>>y by providing bar(y) until they start hunting through the docs.)
>>  
>>
> The problem is that foo(x) returns multiple values, thus there are more 
> bar(y) functions for each x than is practical to list. In other words, x 
> and y are not isomorphic.
>
> To get away from the analogy, what I'm saying is that since there is no 
> standardized description for each problem that each user comes up with 
> intuitively, a user searching for specific functionality can take no 
> shortcuts - he'll still have to search through the short description of 
> each library, for which the alphabetic library list is adequate, and the 
> categorized listing even more so. (The descriptions could be made a bit 
> longer, perhaps. Perhaps.)

Now, now.  Let's not discourage innovation in improving the site and
docs.  If Paul thinks it can be done better, he should try something
and see how it works out.

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