Jeffrey P Kantor | 3 Jan 2008 23:31

DMGeneral - Notes from LSST PM telecon 20070103

Don Sweeney
John Schaefer
Suzanne Jacoby
Chuck Claver
Jeff Kantor
Suzy Dodd
Ted Lavine
David Burke
Tony Tyson
Tm Axelrod

Press Release
SJ - Charles Simonyi will be interviewed by NY Times.  Seems to be slow going right now, but only sent to partners and astro community.  Simonyi may re-issue with business wires.
TT - Dennis Overby from NY Times contacted, and we are on Wired already.  Expect most activity later today and tomorrow.

All Hands Meeting
DS - Need strawman agenda, identify speakers.  Approximately 3 hours per subsystem in plenaries.
CC – The theme is mid-course check on PDR progress.  Identify major PDR milestones for PDR now.
DS – Could do a “dry run” with outside reviewers, but probably too early, need AHM to be work in prep for this.

AAS
DS – Next week, but will still have Thursday meeting next week.  DS wants Ray Plante (AHM) and Andy Connolly (simulations) to participate in these Thursday telecons regularly now.

Budget
DS – Process to finalize is not fully worked out, but we need to move ahead.  Starting with firm commitments, and then have a meeting.  Universities and LCO put money in, all of them put in the first half in 2007, but only 2 gave remaining half in 2008.  We don’t want to pay back 2007, but we will offer to repay 2008 only (for parity), from funds we already have in the bank.
TT – Technically, they could ask for it back, but we are still in fund-raising mode and they need to understand that.  The agreement with the new donors would not permit this.  We need to leverage this new money news for new fund-raising, more people should participate in Thursday meetings.
JS – Too soon to decide, but we don’t intend to use the new funds to pay back those loans.

DM
JK – In slowdown for holidays/vacations. DC2 in “performance improvement” stage.  Will complete in days or weeks, not months.  All nightly pipelines in place (minimal instrument signature removal since we are using LSST pre-cursors).  Participating in AAS next week.  WBS being reworked with responsibility matrix out shortly.  PMCS improvements have been started, User-Defined fields added, resources added (but  more to do).

T&S
VK – Happy about press release.  We have larger versions of images on web, can send to reporters, sent one to SJ already. Rock will start to be removed at CP in January, 2009.

Camera
KG – Wants to ensure press release gets to IN2P3. DOE was hard hit on budget, LSST budget has been affected, makes IN2P3 work more critical.  Sensor RFP will be out shortly, just need final comments for incorporation.  Have wafers from study phase and they will be diced and evaluated in time for PDR.  Will have prototype sensor with Keck funds, but not in time for PDR.  Test dewar is already built at BNL.  Did significant change to Camera section of lsst.org, including filter curves.  No longer have total system throughput shown on lsst.org.

Sys Eng
CC – Working on completing traceability in SysML model, and draft of Sys Eng planning document.  Leaving Sunday for AAS, full week there.  Going to DES CD2 at end of January at Fermilab.  Sent list of suggested changes to lsst.org, including system overview page.

Calibration
DB – Analyzing CT data, use of Y4 looks better (vs Y3), due to efficiency of spectrograph in that band.  Marginally meeting spec with Y4, reproducible at 5 milliMag level summed over 3 nights.  Still checking on systematics, report out by end of next week.  Have approved and scheduled 6 nights on CT with 1.5m and 0.9m (late April and mid-July).  F2F calibration simulations meeting 1/26 – 1/27 at SLAC.  Starting engineering design of dome screen for instrument calibration.  PS failed first light, prototype never left Harvard, Stubbs will measure emission angles/distributions of screen, compute illumination patterns.

IPAC
SD - Met with Robert Lupton, he gave talk about Application Framework.  Zeljko Ivezic gave talk a couple weeks before.  Several people interested in participating, science collaborations and algorithm designs.  Submitted comments to DM R&D plan and owe Applications organization.
_______________________________________________
LSST-data mailing list
LSST-data <at> lsstcorp.org
http://listserv.lsstcorp.org/mailman/listinfo/lsst-data
Jeffrey P Kantor | 4 Jan 2008 16:18

DMGeneral - LSST Funding Press Release

Hello all,

Please see the press release website about exciting funding news for LSST!

Jeff

------ Forwarded Message
From: "Suzanne Jacoby" <sjacoby <at> lsst.org>
Date: Fri, 4 Jan 2008 05:57:05 -0700
To: "jeff kantor" <jkantor <at> lsst.org>, "donald sweeney" <sweeney <at> lsst.org>
Subject: RE: lsst-data

Jeff, definitely yes.  You can direct people to the website http://www.lsst.org/Press/LSSTC_Jan_3_2007.pdf
 
I’m just checking the web for coverage now; Wired has a particularly nice story:
http://www.wired.com/science/space/news/2008/01/internet_telescope
 
Suzanne
 

Suzanne H. Jacoby

Manager for Education & Public Outreach

LSST Corporation

4703 E. Camp Lowell Dr., Suite  253

Tucson, AZ 85712

520-322-8740 (Direct line)

520-881-2626 (LSST main number)

520-881-2627 (Fax)


_______________________________________________
LSST-data mailing list
LSST-data <at> lsstcorp.org
http://listserv.lsstcorp.org/mailman/listinfo/lsst-data
Jeffrey P Kantor | 4 Jan 2008 16:19

DMGeneral - LSST Funding Press Release

Hello all,

Please see the press release website about exciting funding news for LSST!

Jeff

------ Forwarded Message
From: "Suzanne Jacoby" <sjacoby <at> lsst.org>
Date: Fri, 4 Jan 2008 05:57:05 -0700
To: "jeff kantor" <jkantor <at> lsst.org>, "donald sweeney" <sweeney <at> lsst.org>
Subject: RE: lsst-data

Jeff, definitely yes.  You can direct people to the website http://www.lsst.org/Press/LSSTC_Jan_3_2007.pdf
 
I’m just checking the web for coverage now; Wired has a particularly nice story:
http://www.wired.com/science/space/news/2008/01/internet_telescope
 
Suzanne
 

Suzanne H. Jacoby

Manager for Education & Public Outreach

LSST Corporation

4703 E. Camp Lowell Dr., Suite  253

Tucson, AZ 85712

520-322-8740 (Direct line)

520-881-2626 (LSST main number)

520-881-2627 (Fax)



_______________________________________________
LSST-data mailing list
LSST-data <at> lsstcorp.org
http://listserv.lsstcorp.org/mailman/listinfo/lsst-data
Ray Plante | 5 Jan 2008 12:07
Picon

can we tag fw 2.1?

Hi fw-enthusiasts,

DC2 setup would be eased if we cut new releases of all our packages, built 
and deployed in dependency order.  (This stems from the fact that our 
scons utils modifies a package's .table file to require the versions of 
packages it was built against.)  mwi 2.1 is now available and installed on 
the cluster, so I would like to cut fw 2.1 next.  Any reason I cannot 
proceed with this?

thanks,
Ray
Robert Lupton the Good | 5 Jan 2008 14:34
Picon
Favicon
Gravatar

Re: can we tag fw 2.1?

When do you want to do this?  I have some minor changes to Kernel
to support some metaprogrammingish cleanup in imageproc that
should go in some time, but need more testing.   Maybe we should
go with 2.1 now, and be aware that we may need 2.2 soonish --- that
should be pretty easy, right?

Ray Plante wrote:
> This stems from the fact that our
> scons utils modifies a package's .table file to require the  
> versions of
> packages it was built against.

Actually it's the install target (calling eups expand_tablefile) that
does this.  Sorry to be a pedant, but there seems to be some confusion
about what eups does, and scons' responsibilities.

				R
Ray Plante | 5 Jan 2008 15:58
Picon

Re: can we tag fw 2.1?

On Sat, 5 Jan 2008, Robert Lupton the Good wrote:
> When do you want to do this?  I have some minor changes to Kernel
> to support some metaprogrammingish cleanup in imageproc that
> should go in some time, but need more testing.   Maybe we should
> go with 2.1 now, and be aware that we may need 2.2 soonish --- that
> should be pretty easy, right?

I would like to do 2.1 asap, so I would recommend your Kernel changes be 
saved for a later release.  It's pretty easy.

thanks,
Ray
Kian-Tat Lim | 5 Jan 2008 18:38
Picon
Favicon

MWI 2.1

Here's what's in MWI 2.1:

* Fixed #245 (Trace fails on empty messages).
* Implemented new logging framework, including various fixes thereto.
* Added float/double/bool/int64_t to DataProperty::toString().
* Implemented #237 (support for includes in Policy files).
* Persistence via TSV replaces existing rows instead of ignoring new rows.
* Added Citizen::markPersistent() to allow objects to be marked as not
	normally destroyed.  FormatterRegistry is now a Citizen.
* TTrace doesn't force the construction of std::string.
* Fixed static variable initialization in DbAuth and Citizen.
* Added DataProperty::addChildren() to add all children of a source
	DataProperty to a destination DataProperty, replacing any that
	already (uniquely) exist.
* Improved DbStorage::createTableFromTemplate() exception message.

--

-- 
Kian-Tat Lim, LSST Data Management, ktl <at> slac.stanford.edu
Kian-Tat Lim | 5 Jan 2008 19:06
Picon
Favicon

Re: can we tag fw 2.1?

Ray wrote:
> This stems from the fact that our scons utils modifies a package's
> .table file to require the versions of packages it was built against.

Robert wrote:
> Actually it's the install target (calling eups expand_tablefile) that
> does this.  Sorry to be a pedant, but there seems to be some confusion
> about what eups does, and scons' responsibilities.

My question is why we have this call to expandtable at all.  In
particular, expandtable tests logical conditions in the table file
against the setup version at install time, but then they are overwritten
with that version.

It seems to me that explicit versions in a table file should require
exactly that version and no others (which I don't believe is the current
behavior of eups_setup), that no version should indicate that any
version (meeting other packages' requirements) is adequate, and that a
specified logical condition should be adhered to.  eups_setup then needs
to find a set of versions, preferring the versions labelled current,
that meets all package requirements.

--

-- 
Kian-Tat Lim, LSST Data Management, ktl <at> slac.stanford.edu
Michael Dinsmore | 6 Jan 2008 17:56
Picon
Favicon

activity?


I was recently reminded of my interest in the LSST project by news of a contribution by Microsoft:

However, I have looked through my inbox and found no activity on either the internal lsst-team Google mailing list, nor the lsst-data mailing list I subscribed to at lsstcorp.org.  There are no archives for the Google lsst-team list, and I find that I cannot get to the lsst-data mailing list archives, if there are even any--I need a password to access, and I have neither record of my password, nor the ability to have it sent to me as a reminder.

Is there still interest at Google in working with the LSST team?  My interest goes mostly to the data management and analysis aspects of the data, as I believe that a daily generation of 30TB will provide a logistical challenge in it's transport and integration that I am interested in discussing.

--

Michael Dinsmore
dinsmore <at> google.com
Data Center Operative Oregon
_______________________________________________
LSST-data mailing list
LSST-data <at> lsstcorp.org
http://listserv.lsstcorp.org/mailman/listinfo/lsst-data
Ray Plante | 7 Jan 2008 13:01
Picon

update: running DC2 pipelines [DataChallenge]

Hi DC2ers,

David and I have been working on a top level script for routine launching 
of the pipelines as well as cleaning up some of the directories and 
packages.  I'm working on the documentation now; unfortunately, 
I'll have to board a plane (for the AAS meeting) before I can get this 
done and delivered.  I expect to send this out later this morning after I 
arrive in Austin.

I'll share a few status items with you now:

   o  I have created cut-down versions of the input images (thanks to Andy
      and Nicole for their work on the cutter script).  These are found
      under /share/DC2data/D4/LSST.  The original, full-size images are
      under /share/DC2data/D4/CFHT.

   o  I've moved /share/DC2root to the lustre partition (/lustre) from the
      NFS-mounted /lsst partition for improved performance.  Through the
      magic of links, we will continue to use the /share/DC2root path.
      I've started copying the input data over as well; there is enough
      there now for pipeline debugging.

   o  The image processing pipeline is now currently dying in the detection
      stage.  I'll have more info when I return on-line today.

   o  I continue to install newly-tagged releases of our packages.  If
      there is a particular reason I should not do this for your package
      just yet, please let me know.

cheers,
Ray

Gmane