[typo] Re: Seperation between backend and frontend

Tobias Luetke tobias.luetke at gmail.com
Sun Jul 31 14:53:27 EDT 2005

Hi Sascha, 

Thanks for your great email.

Aesthetics is what is pleasing to the eye and or touch. Pleasing or
not can be a very subtile difference.
Take the most aesthetically pleasing consumer device available. The
iPod. Do you remember the uproar about the 8 hour battery you can't
change? The reason why apple didn't put the battery behind an easy to
remove slider is that it would have broken the sleekness of the
design. Yes this is a feature every single other mp3 player or
consumer device of the past 10 years has. In fact it could be argued
that the ipod would never have been as successful as it has if it
would have offered this "feature".
There is a lot of value in leaving out features where many people say
that they are "must have". Fact is that one person's killer feature is
not the killer feature of the next person. In fact
internationalization is not even a common request at all, you are the
only person really pushing for it at this point.
Typo's beauty comes from its simplicity. This simplicity is something
worth protecting but I don't think i do a very good job of it lately.
When typo was still younger i always said we don't need a plugin
interface because changing the code is so obvious. 1.5 KLOCs later....

You need to consider my situation. I get a request a day for easier
migration from MT3 blogs; people want to keep their permalinks ( since
they are supposed to be permanent ). I constantly put these requests
down because i don't want code in typo which deals with just
translating other blogs links into typo's. If we would start with that
we would pollute typo with special case code for 50 different blog

Now internationalization is not a big feature, i have no itch for it.
However it negatively ( thats my opinion ) the entire codebase:
The reason why typo gets a patch a day and boxsom doesn't is that
typo's source is clean. People don't need to care about translating or
re-using already translated strings. Thats stuff people have to battle
in their day jobs and they want to have none of it during their spare

So after we add internationalization the requests will come that each
article can be written in several languages. Every page too.
Categories need to have different names for different locales. People
want to choose the language on the blog by drop down ( impossible with
the current caching ) and so on and so on.

In fact this is a can of worms i don't want to open. I'd accept a
extremely transparent and easy way to translate the frontend in
different languages but there is no chance all that I will agree to
translating the admin interface because it makes changes harder and
the fact that its easy to change is what typo thrives on.

On 7/31/05, Sascha Ebach <se at digitale-wertschoepfung.de> wrote:
> Hi Tobias,
> I wanted to ask you again if you were sure you wouldn't want to have the
> admin section gettextified. IMO *now* is the best time to do it.
> Actually the best time would have been the time where the first
> translatable string was used. Im my experience it will just get harder
> and harder the longer you wait. Also, it is just cumbersome to maintain
> two different set of developing methods. It actually is a violation of
> the DRY principle.
> You said in the channel:
> "Lets just get the frontend translated for now. We can still look in the
>   admin when things are settling down. You might think that its a small
> thing to embedd a string in <%= _(' ') %> but it seriously screws with
> code aesthetics which are one of the best features of typo"
> The argument of aestetics is important. For me the argument of good
> localization is much more important. Any kind of localization will be
> disruptive to your understanding of code aesthetics. What is aesthetics
> anyway? In my view it is making what you have (all the functionality)
> most pleasing by way of molding it into a beautiful form. Something
> cannot really be beautiful if you leave out something important. If you
> think that the Mona Lisa is aesthetically pleasing, what about if you
> leave out her mouth and eyes? The form is not complete and you can't
> really call her beautiful. Transferred to Ruby code you could say that
> one type of string delimiter (') is more aesthetic. It makes the source
> more readable and predictable. But if you leave out the other delimiter
> (") with its interpolation functionality you just loose too much of the
> good stuff.
> If you say "for now" that implies that some time in the future you will
> do it anyway. So when you start today you will already be ready and
> accustomed to the way the localization works. Also, all the other devs
> will have a chance to learn this sooner, when the code base is still
> comparatively small. (Although I must say that it is already quite
> large. It has grown immensely since I last looked). And, let's be
> honest, when will this "settling down" of Typo be? I suspect never. Typo
> is getting more and more successful and popular. The adding of I18n will
> only add to that.
> I see it like that. At the moment lots of people are moaning about the
> i18n support in Rails. Rails was extracted. If Typo will lead the way in
> showing how to do i18n in Rails it will probably get extracted, too.
> This is good, right? Why don't we make the i18n in Typo "look good"
> right now? Or atleast start to do so, because it is going to be an
> incremental change anyway.
> What do you think?
> Sascha

http://www.snowdevil.ca - Snowboards that don't suck
http://typo.leetsoft.com - Open source weblog engine
http://blog.leetsoft.com - Technical weblog

More information about the Typo-list mailing list