[llvm-commits] CVS: llvm-www/pubs/index.html pubs.js

[Nuno Lopes]

>>>>> Log message:
>>>>>
>>>>> Converted all papers from hand-coded HTML to BibTeX-like  
>>>>> Javascript.
>>>>
>>>>
>>>> I don't really agree with this change. This will rule-out non-js  
>>>> aware
>>>> browsers plus some search engines' crawlers.
>>>> If such abstraction is desired, I think it could be done at server-
>>>> side with
>>>> PHP (although I'm slightly biased :)
>>>
>>> Hi Nuno,
>>>
>>> Maintaining the papers list is a major pain (for me, since I am the
>>> only person who seems to add papers).  All the individual papers'
>>> pages are still plain HTML.
>>>
>>> Besides that, I wouldn't be surprised if major search engines had  
>>> some
>>> way to handle this sort of JS.
>>
>> It's true that many search engines support JS, still I know many  
>> guys that disable JS on their browsers..
>> Anyway, I send in attach a simple PHP script that parses the current  
>> pubs.js file format and generates the html automatically. The output  
>> is just an example, and needs some tweaking.
> 
> Hi Nuno,
> 
> I'd really rather not do this.  This sort of thing makes it very  
> difficult to hack on the web page remotely (not connected to the  
> server) and raises maintenance issues (if we ever upgrade PHP or move  
> servers, it adds another thing to worry about).  I think PHP is a  
> great solution for things like the nightly tester (which has to  
> integrate with a database) but for something this simple, I think JS  
> is a great answer.
> 
> -Chris

Ok, ok :)
I just had to defend PHP here, but I won't insist any further.

Nuno