SAMP Web Profile

Jonathan Fay jfay at microsoft.com
Wed Dec 8 10:08:50 PST 2010 

Mark,

Thanks for your hard work on this. After seeing Andy's presentation at ADASS (congratulation BTW Andy on your prize) it was clear that the web apps are getting very rich and interconnected and need to play well with the SAMP desktop applications, but making these tools accessible to the widest client platform usually mean supporting wide reach tools like HTML5/JSCRIPT, especially with the proliferation of devices that don't support Java/Flash/Silverlight. Not that these devices will use SAMP (for now) but why should a developer have to support two code bases (or more) for different clients.

I would be great to see Andy's ASCOT platform support the proposed Web SAMP profile really soon. I know I will take a crack at getting WWT's Web client up on it, maybe in time to show at AAS.

We should probably start a list of links to applications that are going to experiment with the proposed profile with links so we can all test interoperability. A central Web SAMP profile resource page would also be a good idea.

Thanks again for your effort.

Jonathan

-----Original Message-----
From: apps-samp-bounces at ivoa.net [mailto:apps-samp-bounces at ivoa.net] On Behalf Of Mark Taylor
Sent: Wednesday, December 08, 2010 7:32 AM
To: apps-samp at ivoa.net
Subject: SAMP Web Profile

Hi all,

for those of you not in Nara: in the Apps session today I made a presentation, based on work done by myself and Jonathan Fay, and funded by Microsoft Research, proposing a new SAMP profile (alternative to the Standard Profile) for use by web-based clients.
You can find the presentation here:

   http://www.ivoa.net/cgi-bin/twiki/bin/view/IVOA/InterOpDec2010Applications

and an associated web page with some additional explanation, demos, and a draft of the detailed documentation here:

   http://www.star.bristol.ac.uk/~mbt/websamp/

A show of hands during the session indicated that those of the SAMP community present at the meeting were keen to see it taken forward, and moved towards becoming an IVOA standard.
Time was relatively short during the session, so we weren't able to pursue much in the way of discussion of open issues.

For those of you who are in Nara, if anyone would like to take the opportunity of colocation to discuss any such issues face to face I'd be happy to do so.  My guess is that people won't have had time to look at the details enough to make a break-out meeting worthwhile, but if anybody disagrees get in touch and we could maybe do that.  Obviously grab me in coffee breaks etc if you want to during this week, otherwise air thoughts on this list.

Issues that occur to me that might benefit from more discussion are:

  - Security issues (already mentioned by Ray and Bob)
  - Other possibilities for cross-domain workarounds?
  - Details of the hub external URL proxying mechanism
  - Hub API details
  - Choice of port number
  - Plan for how to take the proposal forward

and no doubt some things I haven't thought of.

Regarding the last point, my ideas about where to go from here are roughly these:

 - I will make a public release of JSAMP including implementation of
   a hub and client conforming to the (intial draft of the) Web Profile.
   The hub will be potentially dual profile (Web Profile + Standard
   Profile), but with the Web Profile functionality deactivated by
   default.  In this way most users of the new release won't see the
   new (unstable, possibly insecure) functionality, but developers
   can play with it (e.g. write clients) if they know what they are doing.

 - I will think about tidying up and releasing the JavaScript client
   library that underpins the Web Profile demos on the web page.
   This library works, but is not well written, so needs some
   reworking or rewriting.  I'm not sure yet if I have time to do
   this in the near future.

 - Others (volunteers anyone?) could provide client-side implementations
   in various browser-based languages (JavaScript, either based on my
   JavaScript library or not; Flash; Silverlight; maybe somebody else
   could try to get an unsigned java applet or unsigned JNLP client
   working, which I've tried but can't make work for reasons I don't
   understand).  Thomas has already started!
   It would be very good if we could have another hub implementation,
   which most likely means Luigi - he's said that he's willing to
   implement Web Profile support in SAMPy, but I don't know what
   timescale is feasible for this.
   I encourage sharing of successes and failures on the apps-samp
   mailing list.  Experience using different browser configurations
   will be useful too.

 - With the implmentation experience gained from these activities,
   we can revise and improve as appropriate the draft specification
   that I've drafted.

 - Given that, we work towards the Web Profile becoming defined in a
   new section of the SAMP document, to form SAMP v1.3.  Alternatively
   we could consider it as a separate document referenced from SAMP.
   Again, the timescale for this remains to be seen, but I'd hope
   we can make some progress in the first half of next year.

Mark

--
Mark Taylor   Astronomical Programmer   Physics, Bristol University, UK
m.b.taylor at bris.ac.uk +44-117-928-8776 http://www.star.bris.ac.uk/~mbt/

More information about the apps-samp mailing list