Who else feels that multi-lingual is overlooked by developers?


(Willem) #1

It appears that developers seem to overlook the need for multi lingual audiences all over the place.

Just recently I ran into a problem that if 2 calendars were on the same page using RWML (to automatically show the “other” language version) the calendar stopped working. It caused so much trouble to fix it, that the developer decided to NOT to support it.

Thats logical if hardly anyone uses more than one language. But a HUGE assumption on the part of a developer to think that the majority of needs out there is mono lingual.

I’m wondering if a lot of you don’t stay away from multi lingual NOT because your clients don’t want it, but simply because developers don’t “get” it?


(Jannis from inStacks Software) #2

Fyi: here is another stack targeting multi language, but which focuses on text currently: https://rapidweavercommunity.com/addons/stacks/localizer-stack


(Konstantijn Van Calster) #3

@willemn
Hi willem sad to hear this. I have always got top support from Joost the developer of the RWML stack… very satisfied with his support and RWML stack. Hope you can sort it out, good luck.

Regards


(Willem) #4

Sorry, I was not referring to RWML. I think the RWML stack is great, but other stacks and solutions don’t consider that they need to “play” well with it.

The latest Datesnap 2 version is not able to run with 2 copies of date snap present on the page for instance. It just occurred to me how its mostly an afterthought or not a thought at all.

Mostly, I’m curious how many people actually need multi lingual capacity. If its very few, I’m the first one to concede that there is no case to make it an issue to begin with


(Willem) #5

Thanks for the link @instacks


(Will Woodgate) #6

This is not something I overlook. A large majority of my customers are international. In most instances, my stacks can be used in multiple on the same page. And in themes, I provide options to customise text labels. A prime example being the updated Media theme I am announcing later today.

However it is important to note that not all stacks like to be started ‘hidden’ as the page loads - simply because they may need to fetch calculations or other variables. In such instances you may need to use callbacks to reinitiate a stack, when a link or button is pressed. Again this is something I try to support in my own stacks. If RWML supported callbacks, then things would improve further.

If a large component of your website is multilingual support, then I’d say RapidWeaver is not the best tool to be using in the first place. Multilingual is best done server-side. Platforms like concrete5, Joomla and Drupal really excel at this - including features like auto detection of language (based on browser preferences and IP addresses) and automatically mapping different language versions of pages across the sitemap tree.


(Willem) #7

Very helpful to know, thanks!

The idea of auto detection sounds cool, but in practice can be very annoying as well. If you travel/have business in Europe for instance, you may pass through 4 language changes in a couple of hours of driving. Or visa versa, I’m in my own country, but prefer to speak/use/read another language.

Personally I’m already super annoyed with Google who thinks they can read my mind and keep trying to log me into their local language sites wherever I go. Thats a solution that creates a problem instead of solves it. haha


(Konstantijn Van Calster) #8

Compleet agree with @willwood.
I need to mention also that I have got superb support of @willwood for my Boutique theme with the multilingual RWML stack. But of course not all developers do.


(Willem) #9

@willwood So what I’m hearing is, that developers are all over this, but if someone wants decent multi-lingual you should go elsewhere because… it requires a database to store the content? Meaning RW will never do that? Or no developer is interested in using that approach with RW ? What actually keeps you from doing that?

It would indeed explain my thesis.
I don’t agree that the solution is to go to these other platforms, since they are exactly that: content delivery/blogging type platforms, not website builders. The cost (time and resources for setup and maintenance) for the user that wants a great design and a good content approach is too high.

So in the end, a person wanting to build a great looking website with an effective multi lingual focus is left to struggle with a lacking RW and the inability to scale up to some major platform. So yes. In the cold.


(Rob D) #10

I am using RWML for my multi-lingual site. Generally, it works flawlessly, but not without occasional obstacles.

As Will Woodgate explained, some stacks (especially those using JS) need callbacks in order to “reactivate” their instances “hidden” in language versions other than the one that is presently used. If you change the language version—without those callbacks—you loose the functionality of those stacks. The workaround (and I mean, very sloppy workaround) is to refresh the page in a browser. But if you change your language again, you have to refresh again.

So, I guess, we need to put pressure on developers of multi-lingual stacks to implement the callbacks?

P.S. I agree with you that, in general, the market for multi-lingual customers in RW community is grossly underdeveloped…


(Willem) #11

It would be nice to have a multi lingual “best practice” listing/document somewhere, so all developers can check that against their development plans…


(Chris) #12

Actually RW was never intended to be used with multiple language sites so most of the solutions will always be a kind of “hack”. For example, using javascript to show/hide different elements on the page. Handling multi-language should always be done on the server, else you will also have very bad page load.
That’s the reason why RW is probably not the best solution for bigger multi-language sites.
Setting up WordPress (or other CMS) with multi-language isn’t very hard and once you have done it, it can even be much easier, faster and better to maintain on the long turn.


(Rob D) #13

Actually, my pages load in a blink of an eye – within same language as well as between languages…

I think your approach is a bit outdated. Multi-language platform is here to stay within RW, so, instead of avoiding it, developers should “push the envelope” to achieve better results within. It’s challenging, but all progress ever is.


(Jannis from inStacks Software) #14

Therefore, I designed the Localizer Stack, which allows you to use Total/Easy CMS and Pulse CMS text “blocks” and display them to the user, server side generated, depending on different parameters like browser language, or a PHP session setting.

Sure: it focuses on text only, and partly you are able to switch images also.

Check out https://instacks.com which is more or less translated completely in English and German (and for the parts where no German translation is available, I was just to lazy to do so :slight_smile: ).

Last word: multi language web sites are not only a matter of the tool. It heavily involves content generation in 2 languages, which can be a boring task.


(Rob D) #15

You bet it is, Jannis. But it’s a hell of a lot more accurate that relying on Google Translate…:wink:


(Willem) #16

Thanks for the “other” perspective @instacks. I also agree with @Rovertek.

When the “idea” came about for databaseless CMS, no one got on board fast either. Sure there are downsides to that as well. I for one wanted a database to be abel to “share” that data with a mobile app.
NoSQL also was something no one would have “expected” . My original point was therefore that users and needs determine the feature. Not the available technology or current established practice.


#17

Hi @willemn

Just shared a new beta version of RWML with you in a PM. The new RWML version is also PHP based for most features so hope this already addresses your challenge. Callback features is one of the things I am considering for the future. At the moment we focus on getting thing working in RW 7 Preview in combination with compatibility in terms of upgrading and improving some additional SEO features apart from the server side generated content.


(Will Woodgate) #18

RapidWeaver can only generate static webpages. It has no database and is unlikely to gain database support in the short to medium term. Therefore it is all the more difficult to control what pages or content are served-up to different users.

On the other hand, a database driven website does not have physical pages. When the server receives a request from the user to view a page, the page is dynamically constructed on-the-fly. Therefore you have lots of opportunity to control / manipulate what content the user sees - for example customising it to display in their preferred language. It makes the process of setting up a multilingual website all the more easier and efficient.

There have been database stacks made available in the past (Stack Apps is an example of this). The problem is that they demand extreme amounts of time to develop and much dedication on a customer support front. Let’s be honest; databases are of little interest to the majority of RW users publishing a dozen or so simple webpages. I suspect the kind of price figure a developer would consider reasonable (in respect of the time and skill to develop such addons) would be far beyond what many RW users are familiar with paying. There has to be a business case for developing new addons.

On the contrary, the systems I’ve used look at the browser IP address but also look at the web browser preferences to see what language a user is browsing in. The system can look at more than one metric, so it’s a pretty reliable and intelligent system. On the website itself, a user is presented with an option select menu or flag icons; where they can additionally modify what language they want to view the site in. The website can remember what language a user wishes to view the website in, including on return visits.

That is possibly more to do with the website having an incorrect language attribute - something for the website owner to fix. I have noticed this a lot too. For example, if your page has an EN language attribute and you type ‘lorem ipsum dollar’ above the page fold, Chrome will almost instantly ask to translate the website from Latin to English or vice-versa! Normally if a page has the correct language attribute set that matches the content, translation tools will work properly.

My advice to people wanting to create multilingual websites in RW has been to keep things simple. Just setup two or more subdomains and publish the different language versions of the website into different directories. Provide flags or another ‘clicky’ way for people to jump between the different website homepages.

The RWML stacks are good and I’ve seen people get them work to an acceptable standard. But at the end of the day, having content in multiple languages stored on the same page adds to the webpage weight / inefficiency and is possibly going to cause problems for SEO. And as I touched on in my first post, some stacks do not like being hidden or animated.

Of all the alternative publishing platforms and multilingual systems out there, I consistently find concrete5 to be the best. The system they have in place is pure genius:

Another example:

I use concrete5 for all my bigger client projects. And the company I’m employed by have progressively moved from RapidWeaver / Wordpress towards concrete5 too. Better multilingual support being one of the main factors. The key for it working so well is by starting with a database driven website and opening the system up for actual multilingual copywriters to be able to login, translate content and save the changes.


(Willem) #19

Just to be clear, one of the sites I’m working on has a robust multi page, multi language setup running with lots of content and RWML is doing a FANTASTIC job.

I do not see ANY lag issues or see a need for a database to host the content, except the pages that are handled by Armadillo. We may consider Total CMS if we can figure out an elegant way to handle sharing that data with external apps as well. Either way its a massive project to do any of that. :slight_smile:

We choose Armadillo, because we want to share the content in MySQL with other applications to avoid double entry elsewhere.

There are only 2 issues I’m asking about:
1.How large of a user base is multi Lingual or … would like to be?
2.How many developers of individual (none ML type) stacks actual pay attention to this need?


(Willem) #20

On the “auto detection” side. I wholeheartedly disagree that Google has a clue how to do that. I AM multi lingual and run into their inability to understand language all the time. The concept only works for people who are MONO lingual. It is a classical case misunderstanding how certain people consume content.

There is an app that I use that keeps asking me why I’m reading in Dutch or German, while I have my preferred language set to English. Hahahaha.

I WANT to have all the instructional and navigational information in English NOT in the language I use while browsing. No IP address or “language preference” can solve that. Its just as wrong as Google translate itself. It does not work.