5 ms·
> Internationalization is not a feature, it’s an architecture... It’s not something extra. It doesn’t take longer. It doesn’t require additional engineers.
by crazygringo 8y ago
> Internationalization is not a feature, it’s an architecture... It’s not something extra. It doesn’t take longer. It doesn’t require additional engineers.
That feels like the wrong way to convince people. Having worked with i18n and l10n before, it absolutely is extra, takes longer, and on a very large team can absolutely bump up to an additional engineer.
Producing dynamic strings becomes much more involved and requires more testing. Handling RTL languages and layouts is another layer of things to worry about and test. Giving the translators sufficient context for each string is plenty of work, as is adjusting the design to fit strings that are longer than you'd expected. Etc. etc.
That isn't to say don't do it. It's just yet another cost-benefit calculation taking into account ease-of-use and market size. But pretending the cost isn't there doesn't fool anyone -- it just makes them think you're bad at producing realistic estimates and makes them less likely to trust you generally.
- RcouF1uZ4gsC 8y agoAnd it also causes compromises in the design of the English language version. For example in https://blogs.msdn.microsoft.com/oldnewthing/20030929-02/?p=42363 https://blogs.msdn.microsoft.com/oldnewthing/20030929-02/?p=... Raymond Chen talks about how when a taskbar is docked vertically, the "Start" text disappears because while you can rotate English text to appear vertically, other languages are more problematic. In the same post, he also states that internationalization is why you have "1 folder(s)" "2 folder(s)" instead of "1 folder" and "2 folders".
- Markoff 8y agothis still doesn't work for languages with multiple plurals like Slavic languages, so how you write one word in three or more forms this way? the best thing i come over years it's abbreviating to the core, it doesn't look nice but it works "n fold." will always work unlike changing syntax to "Folders: n" which just can't be used within sentence. only way to make it right would provide different strings for multiple plurals based on value of variable and the best part it's in each language it differ, good luck with that, abbreviation it is...
- blattimwind 8y agoThat's why you need pluralization rules. E.g. https://doc.qt.io/qt-5/i18n-plural-rules.html https://doc.qt.io/qt-5/i18n-plural-rules.html Developer-wise you supply the "n" to the translation function.
- jschwartzi 8y agoAgreed. On the last product I worked on we had a custom UI with no translation/internationalization support and we ended up spending a total of 2 or 3 person-months adding the capability. But then there was also a huge amount of time we had to spend also giving the translators sufficient context to translate the UI because we kept getting inconsistent translations. And there were conflicting regulatory requirements for labeling and units in some places that we had to handle by displaying different assets. We came to the conclusion that we could have spent a month implementing a tool for the translators and probably saved 2 or 3 months of back-and-forth and screenshotting. While it's technically practical to architect for translation and internationalization, it's frequently out of scope for a first version of most products. If you're selling something for the US market, it's difficult to justify bumping your time-to-market to add support for something that you won't need for your first release. And that's not getting in to the additional complexity that it adds to the software system to have that support. It's difficult for me to take someone seriously when they're making statements about internationalization not requiring any additional resources. It just tells me that they have no experience in any of the fields where it does require additional resources. And reading the article, I can see that the "experts" are all from web development companies where there's a framework that exists and you don't have to worry about justifying your use of Software of Unknown Pedigree to an auditor. Or where it doesn't really matter if your translation or internationalization framework doesn't work 100% of the time because it's unlikely to kill someone if a control on your UI is labeled incorrectly.
- cco 8y ago>It's difficult for me to take someone seriously when they're making statements about internationalization not requiring any additional resources. I don't think is the intended meaning of the quote. Based on the rest of his response it sounds like he's saying, "Non-English speaking customers matter just as much to us and as an organization the default should be 'support i18n from the start'." If your default is "we must support languages other than English" then it isn't really "extra work", it is just core functionality.
- hartator 8y ago
- Vinnl 8y agoThe problem with the "all engineers should internalise it into their development process" mindset is that it's held by so many groups. Your architecture should take into account internationalisation, security, accessibility, etc. At a certain point, no single engineer can keep up with it all, and produce working and maintainable code - let alone all of them. Unfortunately, I'm not sure what the solution is here.
- TeMPOraL 8y agoI don't know what the ultimate solution is either, but it seems to me that internalizing more and more things into the development process is one of, if not the main part of growing as a developer. It's not even about being a specialist, just getting an intuition for design features that make it easier to implement i18n/l10n/a11y/security features correctly. And performance. For the love of all that is pretty, do remember to internalize how to write non-wasteful code.
- lumost 8y agoPart of growing as a developer is also knowing what you can do without, if you have 3 months to bring an mvp in front of a single language customer - i18n is the last thing on anyones mind, and delaying for even a day is a waste of precious time. This does not mean that time is the only thing that matters, if you know that at month 6 you're going to launch in spanish - then you better look to avoid major i18n pain points, and think ahead to ensure that between months 3 and 6 there is a viable project to internationalize the product.
- derefr 8y agoIf you’re in the US, maybe. If you’re building an MVP of a dating site for the European market? i18n is feature #0.
- goldfeld 8y agoMaybe an MVP for an European dating platform starts at the minimum of your mother language, or say, the French market? What good is having every language if your MVP marketing budget doesn't stretch enough to warm up the dating scenes of several countries? Better hit big by divide and conquer than try to hug the continent with your legs (to use a Portuguese expression.)
- crgwbr 8y agoYeah, this is very true. I spent 4 years working on, AFAICT, the most widely translated site on the internet, https://www.jw.org https://www.jw.org. It’s currently at 994 languages. Handling RTL and sign languages is an especially large undertaking that can’t just be swept under the estimation padding rug. But, if your CBA allows for it, it’s well worth it.
- jrumbut 8y agoMy i18n/l10n experience is limited but I actually thought it was really positive to get everyone thinking about variable content length and appearance from the outset. A lot of times I get designs that look great with just the right amount of lorem ipsum in there but fall apart when the real content with all its variety and imperfect. I think sometimes having to step back and generalize helps avoid specific problems.
- sbov 8y agoI would love to get a design that works straight away in even just 1 language. Usually I get one that completely falls apart once you add in things like validation messaging or other dynamic elements.