On Fri, Oct 15, 2010 at 10:54 AM, gabouy
<g...@gabouy.com> wrote:
Hi,
I´m an MsC Student working on a mashup related thesis. I´ve been trying the provided EMML samples, along with the reference implementation and I have the following doubts:
1. First of all where should I send technical questions, feedback and concerns? The OMA Interest group doesnt seem to be a place for technical questions, but I could be wrong.
OMA technical alias is
2. Would it be possible to access the RI source codes?
No, RI is provided only as runtime implementation engine.
3. Is it possible with EMML to define a lazy type of loading in a foreach with a nested invoke. I´ve been trying to implement the following type of mashups:
First grab list names from page A.
Second for each name in list retrieved from A, invoke page B, with name as query parameter, and append to result document.
If implemented all server side, in one server loop the response time for the whole mashups is way above the 3 seconds threshold. On my plain java reference implementation I had to resort to lazy triggering an ajax invoke to the nested invoke. Would it be possible to express something like this with EMML?
In EMML Samples, check out
concurrent1.emml
concurrent2.emml
It's basically a parallel-foreach loop. Each loop is spawned in a separate thread, and subsequently fuse statement is used to merge results together.
Let me know if that would work for your use-case.
4. If I want to make a Google Maps type of mashup, where I display some info on a map, what would be the suggested way to do it in EMML? More generally are there means to integrate presentation componentes in EMML?
EMML scripts act primarily as data-services, and are accessible as REST api. So, for client-side mashups you can use any widget/gadget and pull data from EMML using REST endpoints. To facilitate data representation on Maps, you could use EMML annotate statement to geocode address data (geoannotation.emml sample)
There was a bug when dealing with non-UTF-8 charsets (like ISO-8859-1) which has been subsequently fixed, but not yet released. We will get back on how-to work around this.
thanks,
raj