[according to the Web2forDev archives, this response was #30 in a thread indexed under its last response on January 18, 2008]
Re: Consider issues in developing countries
on September 7, 2007
Hi Eddie, all, I'm new on this group (thanks to Tobias for giving me the pointer) and am interested to read the various postings.
I'd like to briefly respond to Eddie's posting and also bring in some thoughts related to earlier posts in this thread by Tim and Steve, and more recently by Louise. (The main focus of my remarks is Africa.)
Eddie poses several questions, the answers to which would involve several factors. One of those that receives comparatively little attention is language. I don't mean language as a "barrier" but language as a way and opportunity to make ICT more relevant and accessible to people's lives, hopes, knowledge, and cultures. The process of doing that in ICT is often referred to as localization.
Tim mentioned language in the context of relevant content (which I consider as part of localization). This is not a new issue, but it is still needing more attention for many languages and their speakers. Localization of content (broadly speaking, translation or creation in relevant languages) is relatively straightforward in the case of static web content, thought there may be issues of fonts & rendering. In the case of Web 2.0 and interactive content, tools for users are key. In all of this, ICT4D and Web2forDev projects need to build bridges with language specialists as well as technicians familiar with multilingual technical issues.
Localization is also a key to full access. For years now we have been referring to access as primarily a technical issue, an extension of connectivity, even as various people - like Louise on this group - and agencies have pointed out that access has several aspects. Physical access is of course basic, but "soft access" may involve, among other things, localization of interfaces, input methods, language tools, and software. In this area, I think there is a lot more that ICT4D projects (and emerging Web2forDev projects) can do, even without getting as far as software localization. This is particularly the case in Africa, where most projects seem to stop at English or French.
Steve mentioned the work of Geekcorps in Mali and the Wikipedia in Bambara, among other things. It is worth mentioning that a Geekcorps volunteer, Kasper Souren, was pivotal in catalyzing some work on the Bambara Wikipedia a few years ago (and has continued to assist work on that and others such as Wolof as a Wikipedia expert). Mali, by way of introduction for those who don't know it, has French (inherited from the colonial period) as an official language, but that is spoken as a second language by a minority of the total population (one estimate I saw - 15% - is probably too low, depending on how one measures language ability. In any event, Bambara (a Manding language very close to Jula, Malinke and Mandingo) is spoken as a first or second language by 80% of the population, and various other indigenous languages are important in different ways. Yet when a USAID-funded project set up several "CLIC" telecenters across the country a few years ago, no thought was given to anything other than French (and English). Whatever the other strengths and virtues of the funders and project personnel, they were apparently clueless on the language dimension. So much so that when I contacted an administrator the reaction was positive, but the only information they had was about some out-of-date legacy fonts for Bambara (which hadn't been installed on any machines) and nothing about Unicode fonts which are of course fundamental to multilingual computing. I don't think anyone decided *not* to consider or accommodate Bambara and other "national languages" (as they are referred to in Mali), but that no one thought about it. So Geekcorps does some interesting work around the edges, and the main ICT4D players pay no heed to national languages, while agencies that could help with the latter are left out of the loop.
This overlooking of African languages in ICT practice is, in Africa at least, the general rule (there is sometimes mention of the topic in some discussions and NICI plans). I have recounted elsewhere how I spoke with Americans connected with an ICT4D project in Senegal who were either dismissive or genuinely puzzled by the notion of accommodating Wolof or other Senegalese languages on computer systems being deployed around the country. Such linguistic narrowness is significant, I think, in terms of =he potential for Web 2.0 and more broadly the potential uses of and access to ICT in multilingual societies.
I took a look at Eddie's blog at http://nyaneba.wordpress.com/(approve sites) and noted a video about a literacy class at ACCASI (Atakwaa Community Center for Agri-info Sharing through ICT) in Ekumfi Atakwaa, Ghana. It was being held in Fante, presumably the first language or local lingua franca of the community. This is no surprise since people learn most quickly in the language(s) most familiar to them. The question is how well the ACCASI effort will support agricultural information in Fante (i.e. will it be localized?), or if it is expected that the information will be conveyed in English only and people will adapt somehow.
This question connects with other issues that go beyond the usual realm of Web 2.0 or ICT4D discussions, such as language standardization and policy. Fante, for instance is very close to Twi Ashanti and Twi Akuapem, but has a separate orthography. Some consider them interintelligible variants of a single Akan language (with probably 9 million L1 & L2 speakers), which itself has a unified orthography. What kind of official support and community use is there for these alternate standardizations?
This is the kind of question that I think ICT4D projects throw their hands up at, and resort to English-only (or French-only), but ironically it is ICT tools that facilitate accommodation of alternate orthographies, and indeed multiple languages. And anyway, a project does not have to get too deeply in any of this to at least assure that the basic (often free) tools for handling local languages are available: beginning with Unicode fonts and supplementary keyboard layouts.
ICT4D programs and localization (L10n) programs in Africa have tended to evolve separately, with the latter receiving a lot less in terms of resources. But that is beginning to change with, among other things, IDRC's support of PanAfrican Localisation efforts (it also funs a PanAsia Networking Localisation project). More communication and collaboraition on these topics is necessary. Fortunately the starting points are relatively easy and often no-cost (making sure systems are fully internationalized, with fonts, keyboards; and where available, installing localized free software).
Ultimately what is needed is localization policy which helps guide ICT4D and, increasingly, Web2forDev projects, and links to both ICT policy and language policy processes. This is pretty much what one sees in many Asian countries. Africa's situation is not Asia's any more than it is that of any other world region, of course, but the elaboration of localization policy seems essential in multilingual countries to make ICT more comprehensible and useful for development needs.
This message has rambled on longer than I intended, but let me add one more point lest I be misunderstood. Localization, of course, does not mean replacing more widely-spoken languages with local languages - the message, and the technology, are additive.
Hope this helps.
Don Osborn, Ph.D. email@example.com +1 202-xxx-xxxx
- Bisharat! A language, technology & development initiative
- Bisharat! Initiative langues - technologie - développement