[ocaml-platform] Unique file names

[Romain Bardou]

Le 05/03/2013 09:32, Daniel Bünzli a écrit :
>
>
> Le mardi, 5 mars 2013 à 08:16, Gabriel Scherer a écrit :
>
>> The "provenance" field originally comes from a design of Fabrice Le
>> Fessant, where a java-like ownership URL, eg. janestreet.core.foo or
>> ocamlpro.opam.blah, would be used *both* as more-unique information in
>> the internal name, and as a directive of how to populate the initial
>> namespace (Janestreet#Core#Foo) to be respected by whatever default
>> namespace construction choice is made (eg. put it in the right
>> directory if we choose recursive directory scanning).
>
> Juste a remark on provenance and java-like URL ownership. I understand companies are keen on these things for self-building pride and branding purposes. However it seems software has a much longer life-span than companies and especially company names. As such I'd prefer them to find a good company-invariant library/namespace name and use that rather than make sources dependent on a name that will have to be renamed once they have been acquired by some other company.
>
> In the same vein it would be absurd to publish my own libraries under ch.erratique as if one changes maintainer, there would be no point in this new maintainer having to keep that name. This would entail that library maintainer change entails source-level change in dependent code which is completely absurd and pointless.

I agree: should a hierarchical naming scheme be used, I would prefer if 
library developpers did not put their brand in the names of their 
libraries. I've always been confused by the Java library names because 
of this. "core" is fine ("jane-street" is not), "stdlib" is fine 
("inria" is not), "batteries" is fine.

If a company name is used, I feel like the library is not free and that 
I should not use it. I don't no why :)

Btw, thanks to Gabriel for the summary (I had read the proposals but 
already had the time to forget them...)

Cheers,

-- 
Romain Bardou