[erlang-questions] modular otp concerns
Tuncer Ayaz
tuncer.ayaz@REDACTED
Tue Feb 18 18:15:46 CET 2014
On Tue, Feb 18, 2014 at 3:32 PM, Vlad Dumitrescu <vladdu55@REDACTED> wrote:
> On Tue, Feb 18, 2014 at 3:27 PM, Steve Vinoski <vinoski@REDACTED> wrote:
>>
>> On Tue, Feb 18, 2014 at 9:11 AM, Vlad Dumitrescu <vladdu55@REDACTED>
>> wrote:
>>>
>>> On Tue, Feb 18, 2014 at 12:48 PM, Tuncer Ayaz <tuncer.ayaz@REDACTED>
>>> wrote:
>>>>
>>>> I know that at least Bjoern-Egil has been investigating the
>>>> possibility of splitting up otp.git into sub repos, and before
>>>> anything is set into stone, I'd like to resolve one concern I
>>>> have.
>>>>
>>> One thought about this: there are several levels of modularity
>>> that could be enabled, but as a first step I think that what could
>>> be separated are the telecom-specific libraries (asn1, cos*,
>>> megaco, diameter). Snmp is on the fence.
>>
>>
>> The enterprise version of Riak, which isn't a telecom-specific
>> system, uses snmp because some customers rely on it for monitoring
>> and alerts. Surely other non-telco apps use snmp as well.
>>
>> public_key currently relies on asn1, so it can't go away just yet
>> either.
>>
>
> Sure, that was only my list of things that could be considered as
> extras. Anyway, some applications are more enterprise-y than others
> and IMHO here is where a line could be drawn.
I don't think any of the libs should be unbundled. I've seen projects
split out components, deal with build/coordination issues, and after
years add more and more stuff back to the core repo. With that said,
if the otp team wants to unbundle libs, and has practical solutions
for the two concerns, it'd be interesting to learn what problem
unbundling will solve. IIRC, Bjoern-Egil has been doing otp releases,
so he must know the pain points.
More information about the erlang-questions
mailing list