[erlang-questions] Erlang Extension Proposals (EEPs)

Per Gustafsson per.gustafsson@REDACTED
Thu Feb 1 13:00:56 CET 2007


Dominic Williams wrote:

>Hi Per, and everyone,
>
>  
>
>>In conjunction with the Erlang User Conference 2006 there
>>was a workshop concerning the future development of
>>Erlang.
>>    
>>
>
>I attended the workshop, and was very pleased about the
>decision to have a process for language evolution. Thanks
>for taking the initiative to get things moving.
>
>  
>
>>1. We need to have some EEP editors
>>2. We need an e-mail address for the EEP editors, I have
>>   suggested eeps@REDACTED
>>3. We need tools to generate html from EEPs. I have altered
>>   some of the python tools to acheive this, but the
>>   resulting script is very brittle and probably won't work
>>   for most EEPs
>>4. We need a subversion repository to keep track of the EEPs
>>
>>The draft for EEP 1 can be found at:
>>
>>http://user.it.uu.se/~pergu/eep/eep-0001.html
>>    
>>
>
>I have read the draft and have several suggestions for
>simplifying the process.
>
>1) I don't see any use for informational and process
>   EEPs. Let's just have plain EEPs.
>  
>
I don't see the problem with having informational and process EEPs. It 
just tells the reader what he should expect.

>2) The suggestion made at the workshop was just to use the
>   erlang-questions mailing list. Why have a eeps@REDACTED
>   address ? Let's keep the process open and simple. EEP
>   candidates can just be posted to erlang-questions.
>
>  
>
Submitting by posting to the mailing lists would probably work well.

>3) It was mentioned at the workshop that requiring an
>   reference implementation might be too strict a
>   requirement. It certainly seems to me that such a
>   requirement rather shuts out the user community from
>   being involved. Requiring a section /about/
>   implementation suggestions or details should be
>   sufficient.
>
>  
>
I think that a reference implementaion should be required before a EEP 
is accepted. A reference implementation is not needed to become an EEP. 
I think it would be hard to accept a "Standards Track" EEP without an 
implementation as it is difficult to evaluate an EEP without an 
implementation.

>4) There is a simple way to avoid completely all the hassle
>   of having a special format, a version repository, using
>   tools to convert to HTML and so on: just hyperlink to the
>   erlang-questions archive. We just need a plain EEP web
>   page, with a table roughly like this:
>
>   N°    Title                          Status
>   ---------------------------------------------------------
>   1     Reified environments           Draft
>                                        Final
>                                        Accepted
>   2     Whatever                       Draft
>   ...
>
>   Each "status" word would simply be hyperlinked to
>   whichever post to erlang-questions had allowed the EEP to
>   achieve that status.
>
>   With such a scheme, several /existing/ posts in the
>   archives could already be given draft status.
>
>  
>
I think that there is a point to require an EEP to conform to a certain 
format and to publish EEPs in a consistent format on erlang.org because 
it makes it easier to refer to EEPs.

>With such a simplifications, all we need to start are the
>members of the committee. I think the open source user
>community should be represented, maybe by voting for
>a certain number of representatives ?
>
>  
>
When we have selected the commitee members most of the other issues are 
pretty minor and if the conversion tools does not work perfectly now the 
sources for the EEPs would still be available.

Per Gustafsson




More information about the erlang-questions mailing list