and I can see that dropping GEARS may be a solution.
BUT we do need a solution which works offline.
I agree that it would be easier to do this as a web-only project,
but our prime target is the disadvantaged world which has dodgy access to the internet.
Phone connections make the internet accessible to almost everyone, even in
developing countries, but these connections are intermittent and very expensive.
I was surprised during a visit to the USA a couple months ago,
that accessing the web on a PDA is prohibitively expensive - 300x the cost in the UK!
AT&T have a monopoly on 3G data at present, and they're milking it.
Of course this won't last long, but I'm sure this kind of situation will exist
for a very long time in developing countries, where monopolies are common.
To make things more difficult, the project has to have an internet component
which integrates with the standalone, because we can't hand out fair use material
on an SD card (or whatever) where it could easily be hacked.
Also, it would be a pity not to use JSword.
So, the big question remains: how do we get PDAs and PCs talking to both JSword and a web browser,
and use the OSIS datasets?
Do you think the GEARS road to solving this is a complete dead end?
David IB
At 13:36 05/10/2009, you wrote:
Hey there
I just thought I'd mention my questions and thoughts about the whole GAE/GEARS thing...
GAE for hosting
It's good because it offers free hosting to a certain extent.
However, it limits us to only one host provider, when actually we could sit the server on any desktop connected to the internet, or any host server running unix...
It ties us to a non-disclosed database which is non-standard, making it probably harder for people to pick up (including me) outside of STEP. (No idea how we would develop locally - but there's going to be a solution there that's for sure)
It is unlikely to work with JSword apparently because it doesn't support Thread (according to the JSword community), also because we'd have to rewrite locator and drivers for it (also according to the JSword community)
If used with Gears, then we've got a non-relational database on the web and a relational database on the client (SQL Lite) making rather hard to organise our data!!!
As a result, we'd probably have to migrate all the OSIS xml stuff in to the Google database, write a OSIS parser and a display for it (when all this is provided with JSword)
Somehow, we'd have to organise our data differently online than offline, since we're dealing with fundemmentally different databases.
I've got no idea whether the non-relational database aspects would support the kind of timeline/articles/etc. data (probably would)
So my question is, why are we so keen on GAE with all its limitations to us? (Don't get me wrong, I'd love to experiment one day with the Cloud!, but it doesn't seem to fulfill our requirements here)
As for browsers and portability, here's how I see it:
The size of the software is likely to be more than a few MBs especially with the amount of data/articles/pictures we're intending to store in it. Are these going to fit locally to mobile devices?
Mobile devices that want access to STEP will have a decent browser.
Mobile devices that have a decent browser are unlikely to need an offline version.
Desktops/Macs/PC/Linux/Windows etc can function with both.
And my question here is, why do we need to use GEARS to have it work locally since mobile devices are going to be able to get onto the net? Who's going to need this?
I just get the impression there are lots of good technologies out there, some of which we can use (GWT, jetty, java db or exist, etc.) and we want to use every single one of them! But at the moment, I get the impression that using GAE and GEARS will add months of development and headaches as to the organisation of our data.
Anyway just my thoughts. I'm happy to be waiting until we decide...
Chris
No comments:
Post a Comment