reddit discussion about Go turns to D again

Andrei Alexandrescu SeeWebsiteForEmail at erdani.org
Sun May 15 17:06:45 PDT 2011


On 05/15/2011 06:10 PM, Jonathan M Davis wrote:
> On 2011-05-15 15:48, Robert Clipsham wrote:
>> On 15/05/2011 23:39, Andrei Alexandrescu wrote:
>>> On 05/15/2011 02:35 PM, Walter Bright wrote:
>>>> On 5/15/2011 7:43 AM, Andrei Alexandrescu wrote:
>>>>> (That's also why we should think of a shorter name instead of
>>>>> parallel_algorithm... and with this parenthesis I instantly commanded
>>>>> the
>>>>> attention of the entire community.)
>>>>
>>>> Leave it as std.parallel_algorithm:
>>>>
>>>> 1. people instantly know what it is
>>>>
>>>> 2. google will index it as "parallel algorithm", exactly what we want.
>>>> URL names carry a lot of weight with google page rank
>>>>
>>>> Calling it "palgorithm" will get us nowhere in terms of visibility.
>>>
>>> Sounds like a good argument.
>>>
>>> Andrei
>>
>> Unfortunately, I'm inclined to agree. While palgorithm saves on typing
>> and avoids the hideous underscore, parallel_algorithm is far easier to
>> find.
>>
>> Argument #2 I don't find as attractive, but it could be a nice way to
>> increase D's usage (not that many people are looking for it -
>> http://www.google.com/trends?q=parallel+algorithm although that's
>> probably not indicative, considering it contains the majority, who
>> aren't programmers).
>>
>> Besides which, it's not like you'll be typing it all the time, if you're
>> doing things that are likely to need parallel computation then that one
>> line of import isn't going to make much difference to the amount of
>> typing you have to do.
>
> If it's a big problem, then that's where alias comes in. However, if
> parallel_algorithm's function names are the same as algorithm's, that could
> result in a lot of name clashing and force you to either use alias or fully
> qualify the package name frequently. If the function signatures are different
> enough though, that won't actually end up being a problem.
>
> - Jonathan M Davis

The function signatures would be identical underlying the fact that 
their actual semantics are identical.

Andrei


More information about the Digitalmars-d mailing list