Re: A GREAT week in India - but LOTS to do!

hi Alan,

Sounds like you did some great work in India. Comments below.

On 02/12/2017 10:25, J. Alan Bird wrote
> Richard and Ralph - If we have needs for similar work around the Nordic 
> languages then Shilpi would be interested in working with us.

It would certainly be good to investigate whether the Nordic languages 
listed at 
https://w3c.github.io/typography/gap-analysis/language-matrix.html are 
correctly represented, and fill in the blanks.

Perhaps we could start with a preliminary review for those languages, 
and then decide whether it's necessary to pull a TF together. For that, 
we'd need some expert(s) who can run through the gap analysis 
questionnaire and provide tentative responses.


> At a meta level she sees us working with Microsoft in 
> two ways.  First if we can arm her with what would be needed to do a gap 
> analysis she would get an intern or two to do that work under her 
> guidance in Microsoft Research.  

>  1. I need to work with Richard to have a presentation / document put
>     together that says HOW we want a Gap Analysis done.  This should
>     include where things would be kept (assume github) and what the
>     review / qa process would be from this team.  I then need to send
>     that to Kalika and Reverie (more on them below)
>  2. WE need to decide how we want to manage the Gap Analysis for Indic
>     Languages.  More on potential conflict below in Reverie notes.

I wouldn't want to see MS doing things separately from the Indic TF we 
are already setting up. I can't see that working.

We could have separate gap analysis documents and lreq documents per 
script, but there are various commonalities across various permutations 
of script (and therefore language) which should be leveraged by having 
all the work done under the one Indic TF umbrella.

I think we also need to be careful not to immediately bite off more than 
we can chew. To do this properly, there's a fair bit of work involved 
per language. I suggest that we start out with a carefully picked 
selection of scripts (there's a suggestion in the charter) and the 
languages we want to tackle related to those. Then working on those we 
can gain experience that will help bring on the others quickly at a 
later date.

For a first draft of the process involved, see 
https://github.com/w3c/typography/wiki/Setting-up-a-Gap-Analysis-Project

*Please take a look at that and let me know what i'm missing.*


> They are extremely reluctant to 
> work too closely with any "big company" as they haven't had great 
> experiences doing that.  This means we need to carefully define what do 
> we give to them vs. what we give to MSFT vs. what we give to e-sahitya.  
> The next steps with them are:
> 
>   * Get the technical package to them
>   * Define who gets what languages and leadership

We really don't want to have competing work going on, and what work we 
do do needs to get buy in from as many experts as possible, and reuse as 
much of the synergetic outcomes as possible - and there should be many 
of those when dealing with Indian languages. So i can't really see a 
scenario that lets us split up the work into discrete TFs.

I don't know what bad experiences they have had before, but perhaps we 
can explain that the initial bulk of work to be done is just to describe 
what does and doesn't work, and how it should work from a 
non-technology-specific pov. There shouldn't be a competitive or 
threatening environment because we're not implementing any technology - 
we're forming a local community that represents their needs to the spec 
writers and application implementers.

Alolita, Abhijit, do you have any insights here?

> What feels like a tactical implementation might be this notion 
> of a "task force" for each language where an in-country individual from 
> e-sahitya, C-DAC, MSFT or Reverie drives the first step Gap Analysis so 
> we know how big of a problem we're trying to solve.

We set up the new Indic charter to be the TF that coordinates all the 
work, and leverages the synergies between the various languages and 
experts in India.  If we have multiple TFs, we'll need multiple 
charters, chairs, and admin overheads, and we'll lose out on the 
potential synergies across indic scripts.

What we certainly CAN do, however, is have separate gap-analysis 
documents and lreq documents for separate scripts, within the one TF 
chaired by Alolita and Abhijit.  Would that work?

> a call on my Wednesday night so it would look something like this:
> 
>   * Richard - 1500

wfm


ri

Received on Monday, 4 December 2017 16:20:56 UTC