Hi David,

happy to see the developers are closely watching the developments in
the community :-)

> - Pootle still has some important initial goals to finish off, such as
> managing permissions, goals etc.

Sounds great.

> - I don't know much about the architecture of your glossary tool. It
> would be nice to find a way to integrate them that is simple. Could you
> describe the architecture (language, database etc), requirements, and
> basic 'how-it-works'?

Right now, the tool is optimized for Georgian projects and the only
thing it requires is PHP. For database, it uses XML files. Each
project has a separate dictionary XML file (like OOo_glossary.xml,
gnome_glossary.xml). These dictionaries include four fields: original
term; term definition in English; Georgian translation of the term;
term status (approved or not approved).

The users are registered manually. Once a user is registered, a
separate XML for each dictionary is created for this user (like
OOo_nino.xml and gnome_nino.xml if the username is Nino). The fields
in user files are: Georgian Translation; Comments; Status (closed or
open).

When a user logs in and chooses a dictionary, he can browse the terms
(defaulting to 50 terms per page, but user-configurable). The terms
are listed as a table, with columns listing Term ID, English term,
Term definition, Georgian translation, Status and two columns for each
user (translation/comments). The user columns can be hidden by user.
If a user is logged in, his/her columns comes first after the status
column and represents two editable fields where the user can input own
translations and comments. When a user has filled in the fields with
his/her translations and comments, he can click on NEXT or SAVE and
the database is updated, and the system lists another 50 terms (or, as
much as a user has specified).

The converter to convert the user files to PO or TMX (or other open
format) can easily be added to the system.

If the tool was to be integrated in another product, it would be
better to use a real database rather than XML files and have it
convert to and from various formats. If the product to be integrated
in Pootle, than I guess it should be ported to python (correct me if I
am wrong).

> - Translation Memory is also important. It's not neccessarily easy to
> just pull the translation memory out of other programs. I have my own
> ideas about how to do this.

I agree. And it will not be easy to integrate in an online tool. But
as soon as it is done, I will probably be the first person using the
tool for great many l10 projects :-)

Do you have a rough date in mind when this should be available? Be it
a month, six months or a year?

While the product is indeed very useful to many projects the way it is
now, I think we will have to wait for TM functionality before the tool
can be used for OOo l10n.

I think it would be nice to decide which way to go now: Try to embed
the glossary management tool in Pootle soon and have the whole product
embedded to Native-Lang pages, wait for the TM functionality while
integrating the glossary management to the tool and embedding the
product as soon as TM and Glossary functions are fully integrated, or
try to embed the glossary management on NL sites first and later also
embed Pootle, as soon as it features TM functionality.

> If a third-party tool is integrated the
> architecture discussion would need to happen as for the Glossary tool.

Right. But when talking about the glossary tool I developed, I will
put it under GPL and we can just use the code in Pootle (so it will no
longer be a third-party tool). Pootle is also released under GPL,
right?

While I was not able to install the product yet (btw, are there any
instructions how it can be installed without root privileges?), would
you mind stating the features Pootle is offering at present? It would
enable us to decide whether it would be best to integrate the tool
directly or wait until some other features are added.

--
Regards,
Aiet

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to