hacking cmsms for multilanguage

[DeMZed]

Hi !

I've just started to try adding multilanguage support, but It's not as simple as I thought, first of all because there is no comments at all in the code !

But there's something already hard coded in the code. Let's take example of addcontent.php : if i'm not wrong "content_en" is a string stored in a column of the database and hard coded in the code ?

Thx
PS : I try do add multilanguage myself because what I want to propose another way for multilanguage project ...

[cyberman]

Have you seen this solution?

http://dev.cmsmadesimple.org/projects/multilang/

[DeMZed]

Have you seen this solution?

http://dev.cmsmadesimple.org/projects/multilang/

Yes I've seen it ... It's not really clear in my first post but

PS : I try do add multilanguage myself because what I want to propose another way for multilanguage project ...

But I've also seen that the database model must be modified (add columns) to support this conception... It's a real mistake !
What application have as much number of languages as columns in some tables ??

I try to hack CMSMS in order to use several lines in a table with data keys like "content_fr_FR" on "content_en_US'" ... It is a better way of conception, and furthermore, the code I read make me think CMSMS multilanguage is not developped in the way that CMSMS simple had let the door open to support multilanguage...

[Pierre M.]

Hello DeMZed and all,

I hope I understand you well.

Assuming
-you are thinking about the multilang implementation for CMSms,
-you are noting monolanguage is hardcoded today (only one content flavor, stored in one column),
-you would like each content to be tagged by language and allow every content to be available in any number of languages (allowed by storage in rows),
then I'm happy a nice multilanguage implementation is at least being designed. Good work

Pierre M.

[DeMZed]

Hello DeMZed and all,

I hope I understand you well.

Assuming
-you are thinking about the multilang implementation for CMSms,
-you are noting monolanguage is hardcoded today (only one content flavor, stored in one column),
-you would like each content to be tagged by language and allow every content to be available in any number of languages (allowed by storage in rows),
then I'm happy a nice multilanguage implementation is at least being designed. Good work

Pierre M.

You assume well and thanks for the link

After several hours, it seems that multilanguage hard coding is very implented in the code...
RonnyK has certanily understood this before me, and thus, his solution is the cheapest.

I would like CMSms to be multilingual (you can search for my commitment to this in these forums). This is why I like what katon and jsmonzani have made so far. I'd like to stress the "single negociated tree" design pattern because I think it is best of all worlds :
-Pretty URLs aren't touched. You still make the rewriting you want and have the URL you want.
-TCN is automatic and user friendly (and your browser is allready configured for it - all browsers, in fact).
-Universality : no need for cookie hacking, just actual HTTP headers.
-It keeps CMSms simplicity : the templates are not modified. CMSms just serves the requested flavour of the same content bloc. In the admin, we can add as many flavours as we like, just tagging them with "en", "fr", "es", "it", "de", "ru"...
-As it is HTTP implemented, it is webserver cache friendly.

I'd like it be done the web way, the simple way, the CMSms way.

It's a great idea, that my way of think ...

katon once began the work. Now jsmonzani has made steps to resurect it. I'm respectfully suggesting to work with the devs at the core level rather than at the templating hack level. I wish fun coding days to all people involved.

Yeah, sure, but time is money even in Open Source world ! That's why, like others, I try to quickly do this...
But even if I master PHP, breaking down CMSMS to input multilanguage support is an long and entire task which needs perfect knowing of the application...

I mean : hacking CMS MS to introduce "single negociated tree" and all the good design rules we like, it hurts too much CMS design to be kept as a hack : it MUST be a coordinate task,  not done with a Lucky-Luke-Lonesome-Developper style !

[Pierre M.]

Hello again,

I'd like to stress the "single negociated tree" design pattern because I think it is best of all worlds :
(...)

It's a great idea, that my way of think ...

I'm respectfully suggesting to work with the devs at the core level rather than at the templating hack level. I wish fun coding days to all people involved.

Yeah, sure, but time is money even in Open Source world ! That's why, like others, I try to quickly do this...
But even if I master PHP, breaking down CMSMS to input multilanguage support is an long and entire task which needs perfect knowing of the application...

I mean : hacking CMS MS to introduce "single negociated tree" and all the good design rules we like, it hurts too much CMS design to be kept as a hack : it MUST be a coordinate task,  not done with a Lucky-Luke-Lonesome-Developper style !

I like Lucky-Luke
I agree. This is good news.
Just my two cents about the needed coordination : multilang is today in the roadmap for 2.0 and as far as I know 2.0 is very different from 1.x (because of ORM and other improvments).

Let us be optimistic : if 2.0 improvments isolate things better, the content fetching is going to be more isolated and can be patched (fetch this language flavor) without side effects.

And of course nobody here masters time (or 2.0 would already be released

Pierre M.

[DeMZed]

Just my two cents about the needed coordination : multilang is today in the roadmap for 2.0 and as far as I know 2.0 is very different from 1.x (because of ORM and other improvments).

Let us be optimistic : if 2.0 improvments isolate things better, the content fetching is going to be more isolated and can be patched (fetch this language flavor) without side effects.

Well, that is a good news ONLY if it will be possible to move from 1.x to 2.0 ... My own experience tells me it's would not so simple...

And of course nobody here masters time (or 2.0 would already be released

Well, "Toutes les bonnes choses arrivent à temps" !

Oupps, automatic translaton failed

[SN8P]

well, what append to that project?