Undolog.com

As a stand-alone application developer, and especially as a web developer, I often face the problem of locating and supporting multiple "languages". The majority of desktop applications have a major advantage compared to a website and a blog in particular. There are now many applications, including free and / or open-source, easily support multiple languages, including those completely indecipherable, such as Chinese, Cyrillic, etc ... All this shows the presence of software tools, ready, able to offer an environment specializing in the management of different languages. The same operating systems, in fact, from Windows to MacOS, arriving on our desks provided with good support for localization, which is not reduced to mere replacement is of the "characters", but it also works on the direction of writing (see Arabic or Chinese - to right to left or vertically). The operating systems are able to provide that minimum of layer Kernel on which then are laid the various tools present in the market.

Desktop Applications

The situation of Desktop applications and very different from the Web, as we shall see later. The applications that we install on our computer, in fact, enjoy a host of privileges absent in a setting "open" as it's the Internet:

• Normally you get the localized version of a software, if available. Sometimes the same software to incorporate all the languages ​​available to them (international version). The user normally syncs with the language of the operating system that installs the software, or at least tends to do this (if you are Italian will tend to have the OS software in Italian Italian, and I will try ...).
• The software installed on our computer is basically used by us or, at most, by a person who speaks the same language.
• More importantly, the Desktop - with very few exceptions - does not change its contents and, if it does, we ourselves generate.

The Web

The multilingual web sites (or out of monolingual and multilingual become later) suffer from a complex multi-content management. This results in the analysis phase, nell'evincere glance what sections of the site that maggiormanete may suffer from problems of localization to the online (see news, news / press releases), so choose in advance the best technology to manage different localized content.

We can imagine two major models for the location of a web site:

1) Duplication of the structure and content based on different domain names (www.miodominio.fr, www.miodominio.it, ...) and, therefore, point to files and independent trees.

2) Unique structure dynamics, which locates the pages based on the properties of your browser and / or user choice.

The choice of option 1 (expensive in terms of maintenance) ridindante to programmatically (HTML / Javascript / CSS / PHP ...) and any database. Not for the content, in any case, however, should be translated and provided (if I give them to different servers instead of on the same server lovoro translation was done anyway ...). This solution has the advantage of making the tree language independent: for example I have 2 press releases for the Italian, but only 1 for English ... or news for French and 10 for the Spanish ... and so on. The disadvantage lies in the redundancy of the code and in the independence, in fact, of the single site localized related to the other. Although there are various techniques to "share" some cross-domain functionality, thus limiting this independence, a possible technical update on a "site-language" must be "copied" on all the other n "-language sites"!

The option 2 provides a structure Web / unique logic, where the text (and any graphic elements to be localized), is generated runtime on the basis of a series of information. In essence, the Web site but the content is unique - localized - are many. The important point, however, is that in a configuration like this, the system assumes the existence of content in every possible language that "could happen"! The contents, in practice, especially those added over time (news, press releases and / or press releases) must go hand in hand. If "insert" in Italian I have a News, simultaneously, to insert all the other languages ​​supported! Alternatively, and imoprtante here's the thing, you must configure the system so that "covering" the event of a discrepancy between a localized content and more!

Both are articulate and carries with it advantages and disadvantages.

Blogs and user-generated content

If we move from a "simple" Web site to a blog, for example, the situation is nothing short of dramatic. There Plugin (for WordPress, for example) that allow you to manage the insertion of posts in different languages. Basically when you write a post you may decide to immediately write its English version and / or francecse and / or German, etc ... This solution, in my opinion, really solves very little! First of all, the entire interface of the Blog (eg navigation) is not localized comuqnue, everything is concentrated in the Post. Second, the languages ​​supported are decided by the author which, excluding polyglots, makes the job of writing a post a real company.

But perhaps most important point, in the era of Web 2.0 is all user generated content, today, more than ever, the "navigator" is an active part in the consultation of the Web Anyone who can leave their comments and more generally, to contribute in fact to the generation of content. In this scenario it is clear that, having no information on possible future contributions of a user, no preventive system of localization can be effective.
However, in the era of automation, it is incorrect to think that it is the machine to perform the task of translation. The translation systems become increasingly sophisticated over time and are the only real answer to a simple and immediate translation. Easy for people to build the website / blog, and simple for the user.

On this blog, for example, I introduced a Plugin for WordPress ( Global Translator ) able to exploit the simlutanea translation service offered by Google (but there are many others). The translation, though not perfect, can also translate the comments left by readers! A key step in the translation of the future!

Systems based on artificial intelligence and self-learning, ensure - sooner or later - translations not unlike those "human" or at least more widely acceptable than those available - the vast majority of us - today. Moreover, the tagging of HTML code, will be able to intervene even where the operator is forced to surrender (at least for now): as in the case of images that contain text or in the guide within a text to be translated.

This article was useful to you?:

Loading ...

• 

• Tweet