 |
 |
On Wed, 28 May 2003, Malcolm J. Currie wrote: > > I take it that that the "web enabling" of Starlink apps is covered as part > > The E-SO bid doesn't make the point that ORAC-DR can already and cdoes > access web-based information, such as catalogues. Thus one aspect of > e-science can be addressed before the Starlink applications are `web > enabled'. Good point. catalogue access is already in. > > > - You have missed IRIS-2 from the list of ORAC-DR support instruments. > > Yes, and UIST is more than the IFU. There is also a proof-of-concept > NIRI (Gemini) implementation akin to the GMOS one. > I think the current draft AG-II may over emphasise the Gemini angle. > > - You probably should point out that this work is complementary to the > > ESO pipelines since ESO are not supporting user pipelines, just pipelines > > to monitor data quality. > > I thought it was, but it doesn't harm to repat the point phrased in a > different fashion to drum it home. Looking at the eclipse web page it seems to me that ECLIPSE could be called from oracdr for certain operations if that would make it more ESO friendly. > On the Dalton point, we could say that in the consultation period that > we are mindful not to duplicate "grid-enabled" packages being written > for specific instruments. I don't think that they're so complex that we > can't manage. ORAC-DIR may not be as efficient in cpu cycles because it > may take a few steps of atomic applications rather than invoking a > bespoke task to perform the algorithm in one go, but we can save > programming time. The modular nature also makes it more amenable to > modification by astronomers. ORAC-DR does work on a variety of > instruments, some of which have multiple modes, and we've coped already. > We've adapted existing recipes to new instruments. These facts > demonstrate that we can develop ORAC-DR for complex instruments, if > we need to. > > > - including letters of support from UK users of ESO data would be helpful > > I think but you might be running out of time. > > I did go through the couple of years UKIRT observer reports to find > quotable comments about ORAC-DR. If it's right ethically, as these > confidential to JAC, we can mail these people asking for permission to > quote them. What do you reckon Tim? Andy agrees that you can mail people who have given positive ORAC-DR comments in the past. > > I have already suggested asking Mark Birkinshaw, Chairman of the AAO > Board, given the favourable comments re. ORAC-DR introduction at the > AAT. John Davies can give a comparison of ORAC-DR versus his experience > at the VLT. Is there enough time for these people to respond? > > > One of the key benefits is > > making life easy for UK users so that they can make the most out of our > > ESO subscription. I think the benefits paragraph should be fleshed out > > some more. Something about large demand from users, no support from ESO > > etc etc. > > One of the main thrusts of the PPARC plan is to maximise the benefits of > participation in ESO and Gemini. So yes we should push this for all its > worth. Astronomers concentrating on science not software... etc. > I see no-one is asking for money to put Gemini fully into ORAC-DR.... :-) > Our work may also benefit future instrument builders, who don't have to > start from scratch, posibly leading to a saving in ESO's development > budget. (I don't know how new instruments are funded in ESO.) Nor do I. > > > - Have ESO expressed an interest in your ISAAC pipeline? Is it to be > > publicised in the ESO newsletter? > > Not directly, but Derek told me that some people at ESO were impressed. > I already decided we should have a piece in the "ESO Messenger" once the > pipeline has been beta-tested. Besides showing how wonderful ORAC-DR > is, it can address some of the problems, which, for example, may lead to > improvements to the headers as Tim suggests. > > I'm also going to ask for a link on the ISAAC reduction page. The > question of where isn't clear yet but it should have some overview, > links to the Software Store to obtain ORAC-DR and to the various SUNs. > BTW Tim the ORAC-DR home page is a little out of date. > Yes. Hmm. Well. Talk to Brad :-) > > - The one staff year sounds like the work required simply to generate a > > wide range of recipes for ESO data. Actual supporting the Starlink > > applications will be difficult for that person. Where is the money coming > > from for that? > > Personally, I don't think one SY/y on the wide range of recipes is > sufficient in the first year. To make an impact on more than one > multi-mode instrument instrument, even just considering what's already > in the ESO Archive, let alone the new instruments coming. We must have > the main few instruments supported quickly to become the standard for UK > reduction of ESO data. Speaking personally, while it's fine for my > Annual Report being a vital cog in the machinery, it's important to > spread the expertise in ORAC-DR amongst the Starlink team. This can > also lead to further ideas and improvements for ORAC-DR. > If this is the only source of funding for starlink classic then at least 2 people should be funded from this since it sounds like there is no way for one person to do all the ESO recipes *and* support the appplications and add new state-of-the-art algorithms. > The work plan should mention that we wish to improve the documentation > for astronomers to adapt the recipes. This includes the Programmer > Manual. Is xoracdr documented outside of itself? Not as such. xoracdr is mentioned in passing but nothing more. > > > - Are ESO amenable to tweaking their FITS headers (eg by putting in a DR > > recipe or the dispersion) to make life easier for the pipeline? > > My contacts so far would suggest otherwise. I've made a number of > suggestions, and I get no feedback. No even a "thank you for your > comments, and we'll consider your suggestions". I've not been impressed Maybe you should arrange a face to face meeting. Surely it is useful for Starlink programmers to at least meet with ESO programmers. Spoke to Andy about Co-I status: he is happy to be put forward as a supportive community member but the current slant of the proposal is such that JAC as a CoI would seem bizarre (especially given that the current 1 man year is focussed entirely on ESO recipes without regard to algorithm enhancements that may benefit the JAC). -- Tim Jenness JAC software http://www.jach.hawaii.edu/~timj