Re: [perldoc2-developers] [RFC] Platform Specification Vs. 0.1

[Robert 'p. S. <rs...@47...>]

herbert breunung said:

> #####################
> - interface(s)
> #####################
>
> so we do it as webapp not desktop, that would be fine since im also
> involved in several other projects and a tranlation helper is something
> totally different than the pod viewer i want to build anyway.

Hmm. I'd propose we factor the domain logic out of the webapplication. So
that the webface is only the interface between checkins, checkouts,
merges, stage changes etc. This would allow an easy implementation of
something along WebServices, XMLRPC, REST, etc.

> #####################
> - people
> #####################
>
> so we have admins that create and maintain the platform
> and
>
> translators that can be assigned to a whole language a document or as
> reviewer
> more seems to complex for me

Why is it too complex? And what parts exactly? I'd rather think it
shouldn't be too simple, as we have different target audiences and
shouldn't build up barriers.

> even if it should be
> for *ANYONE*
> should be as easy as possible to make comments on any translated passag=
e,
> that can be
> considered later by the assigned translators.

That's for some reason an obviously good idea. The commenting possibility
I mean. As said, I'd differentiate between comments and errata (the latte=
r
being kind of advanced comments). I thought about a ticketized system
because of the history effect and the minimization of duplicate efforts.

> #####################
> Multiple formats
> #####################
>
> we are perl hacker with CPAn on our side so converter tools should be
> least problem.

Full Ack. All the codebase are belong to us. :)

> #####################
> Meta information
> #####################
>
> thats no enough, take me for instace, es i wrote my czech is sophistica=
ted
> but containes
> some terrible orthographical flaws (some people here claim that is also
> true for my german *g* ). such infos i like to atach on a doc that an
> revisor don't have to search for flaws i know there are at certain word=
s
> or word group.
> so only mark a block or whole sentences as fuzzy would'nt do it IMO.

Yea, but it seems easier for translators to reject complete blocks or
paragraphs, but _with comments and reasoning_. As sometimes it's not just
a fuzzy word, but it's context, the sentence around it, and other things
which might be reconsidered to make it more fitting.

> in much i agree with robert

Always a good thing! *scnr* ;)

gr.,
Robert

--=20
# Robert 'phaylon' Sedlacek
# Perl 5/Catalyst Developer in Hamburg, Germany
{ EMail =3D> ' rs...@47... ', Web =3D> ' http://474.at ' }