There's no magic solution. Just work, hard, on the ground work.
For some languages, industrious people from the ethnic group that speaks that language made their own writing systems, which were either partly based on existing foreign systems or created from scratch. Examples from the last couple of centuries include Cherokee, N'Ko, Santali, Vai, and Ho, and there are others. From what I've read about them, they were created by self-taught people who managed to figure out the phonetics of their own languages with little or no formal training in European-style academic linguistics. The creator of the N'Ko writing system Solomana Kante was subsequently praised by European academics as someone who managed to describe the phonetics of the different regional varieties of his language with a well-matching unified writing system, and there are similar evaluations of the other people who created the alphabets I mentioned above.
For many other languages, the writing systems were created by foreign religious missionaries or political functionaries, who also happened to have some understanding of language. It worked better in some cases, and less well in others. When I say "better", I mean that the people who actually speak the language managed to learn it and establish the use of that writing system for elementary literacy education, recording ancestral stories and local knowledge, publishing newspapers and books, personal writing (emails, shopping lists, greeting cards), government and business, and so on. When I say "less well", I mean that little was produced in that writing system other than a translation of the Bible or the Quran.
What should be done? A brand new writing system, or an orthography that is based on an existing one? There's no one answer. Using a Latin-based alphabet has obvious advantages: it's available in
computer keyboards and printing houses everywhere, and a lot of people
are familiar with it. But for some languages other alphabets worked
better for establishing schools, so it doesn't have to be the end-all,
only option. The only real answer is "whatever works". It's a very generic and circular answer, but that's just how it is. Different things worked for different languages in history.
I am not opposed in principle to the hosting on Wikimedia sites of content in languages that have a completely new writing system, whether based on an existing writing system (such as Latin, Cyrillic, Arabic, or Devanagari) or a brand new one. There are some practical considerations with this however:
1. If it's a brand new system, which is not in Unicode yet, it will be technically difficult.
2. If a Wikimedia project is the first place whether a new orthography is used, this may be going against the existing Language committee's principle of "not creating new linguistic entities". I am a member or the committee, and I support this principle. However, I'm willing to be flexible whenever the people involved somehow prove that they are qualified and sincere. (I am writing this only on behalf of myself and not the whole committee. Other Langcom members may have a different opinion.)
3. On which project would such content go? Definitely not Wikipedia in any language. Wikisource may work, although Wikisource till now has been a place for hosting already-published works. Perhaps for new languages Wikisource could become more flexible, or a brand new wiki project could be created.
--
Amir Elisha Aharoni · אָמִיר אֱלִישָׁע אַהֲרוֹנִי
http://aharoni.wordpress.com“We're living in pieces,
I want to live in peace.” – T. Moore