Improve search in digitalmars.com/d (Was: foreach)

[Jesse Phillips]

On Tue, 25 Nov 2008 00:40:41 -0200, Ary Borenszweig wrote:

> Ary Borenszweig escribió:
>> Sam S E escribió:
>>> Jarrett Billingsley Wrote:
>>>
>>>> On Sun, Nov 23, 2008 at 10:48 AM, Sam S E <asdf at mailinator.com>
>>>> wrote:
>>>>> Jarrett Billingsley Wrote:
>>>>>
>>>>>> On Sat, Nov 22, 2008 at 11:40 PM, Sam S E
>>>>>> <eisenstat.aa at sympatioco.ca> wrote:
>>>>>>> Does foreach use delegates? Isn't that unnecessary overhead? --Sam
>>>>>> It does use delegates, for iterating over most types.  When
>>>>>> iterating over arrays, the compiler turns it into a sort of for
>>>>>> loop instead.
>>>>>>
>>>>>> Is it unnecessary overhead?  It's not always as fast as it could
>>>>>> be, but unless someone can figure out some other way of
>>>>>> implementing it, it's pretty much the best we can get.
>>>>>>
>>>>>> How about iterator objects, like in C++ or Java?  Are they
>>>>>> unnecessary overhead?  ;)
>>>>> Why not just use a normal for loop; wouldn't it be almost as simple
>>>>> as token substitution? By 'most types,' do you mean associative
>>>>> arrays or am I forgetting something? As a mainly C(++) programmer, I
>>>>> don't use iterators when I don't need to. I don't even use classes
>>>>> when I don't need to.
>>>> How do you use a for loop to iterate over an associative array whose
>>>> implementation is hidden, or a binary tree, or any arbitrary
>>>> container, or a sequence or words in a file, or the zipped contents
>>>> of two lists, or...
>>>>
>>>> The point of foreach isn't performance, it's flexibility and
>>>> abstraction.  As long as you can make an opApply or function which
>>>> takes a delegate, you can use the foreach loop with it.  Not
>>>> everything is an array.
>>>
>>> Where are the docs on opApply?
>> 
>> Search opApply in http://www.digitalmars.com/d/1.0/statement.html
>> 
>> Can't the search in digitalmars.com/d be improved? Searching for
>> "opApply" you get ten results which are just questions in the
>> newsgroups, and just in the second page you can see the result you
>> probably are looking for.
>> 
>> I suggest improving it in this way:
>> 1. Have a map of obvious keywords that people will look for, to precise
>> urls. Some examples are: delegate, function, override, virtual, class,
>> struct, union, interface, pointer, op*, operator overloading,
>> inheritance, final, const, invariant, etc. 2. If the search is exactly
>> one of the keywords, use the map above. 3. Else, use Google Search
>> Engine.
>> 
>> The list might be big, but it can be done if the community helps, and I
>> think it will improve a lot the usability of the site.
>> 
>> Actually, the map can be [keyword -> list of urls]. For example, for
>> invariant it should point the invariant attribute as well as invariant
>> classes.
>> 
>> And don't point me to
>> 
>> http://www.wikiservice.at/wiki4d/wiki.cgi?LanguageSpecification/
KeywordIndex
>> 
>> 
>> I know it exists, but people how enter the D site (and the above is not
>> the D site) will use the search box. Who uses indexes these days when
>> you can search and find what you want?
> 
> Umm... I didn't mean to sound rude or anything. Unfortunately, when
> someone doesn't speak in its native language, he tends to sound rude.
> 
> Does anyone else have problems with the search box in digitalmars.com/d
> ?

I don't think this really sounded terribly rude. I do agree that 
searching the site can be annoying when top results are always from the 
newsgroup. Maybe an option to search digitalmars/d and one to search 
digitalmars/d/archive (don't know if that could be customized).