[Swift-devel] CCA-11 Swift Poster Visitor's Questions

[Daniel S. Katz]

On Apr 14, 2011, at 7:04 AM, Michael Wilde wrote:

> That looks good for a talk, Dan. The advantage there is that the speaker can point out the topics and describe the code.
> 
> The challenge on a poster is that no one reads a lot of fine print code. So taking code similar to whats on your talk slides, and "magnifying" the main statements to a much larger font, may accomplish the desired effect.
> 
> It would be good to continue to hone a set of talks, posters, and the online material so that we get the message across in a way that shows the strengths of Swift. Given enough space, we can show the inevitably messy details of programming real cases, while keeping the details in perspective.
> 
> All that said, I want to thank Ketan and Justin immensely for putting the whole CCA11 effort together - abstract, cloud runs, plots, poster, and talk - with very little help from me and in a very short timeframe along with many competing responsibilities.  That was very nice work!!!

Yes, I agree!!

> 
> - Mike
> 
> 
> ----- Original Message -----
>> In the Montage presentation I have, I include the code (from Jon) that
>> runs a simple version of Montage, and one of the wrapper examples, as
>> follows. This seems to me to be sufficiently complete for a talk. I
>> use two slides, one for the code, and one for the wrapper.
>> 
>> 
>> 
>> 
>> import {…}
>> 
>> MosaicData header <"header.hdr">;
>> Table images_tbl <"images.tbl">;
>> Image mosaic <"final/mosaic.fits">;
>> Image projected_images[];
>> 
>> Image raw_image_files[] <filesys_mapper; location = "raw_dir", suffix
>> = ".fits">;
>> projected_images = mProjectBatch( raw_image_files, header );
>> images_tbl = mImgtbl( projected_images );
>> mosaic = mAdd( projected_images, images_tbl, header );
>> 
>> 
>> ( Image proj_imgs[] ) mProjectBatch( Image raw_imgs[], MosaicData hdr
>> )
>> {
>> foreach img, i in raw_imgs
>> {
>> Image proj_img <regexp_mapper; source = @img, match = ".*\\/(.*)",
>> transform = "proj_dir/proj_\\1">;
>> proj_img = mProject( img, hdr );
>> proj_imgs[ i ] = proj_img;
>> }
>> }
>> 
>> app ( Image proj_img ) mProject( Image raw_img, MosaicData hdr )
>> {
>> mProject "-X" @raw_img @proj_img @hdr;
>> }
>> 
>> 
>> 
>> On Apr 14, 2011, at 6:41 AM, Michael Wilde wrote:
>> 
>>> That seems a valid criticism that we should address. In this case
>>> the entire script is fairly compact, so we could put the whole
>>> script on the poster and just enlarge the highlights shown here (and
>>> the mappings as well if desired).
>>> 
>>> We should move these to a new Case Studies section of the Swift web
>>> that we have long proposed.
>>> 
>>> The format can be a more visually interesting version of the
>>> annotated script walkthroughs that we put in the Parco paper. (By
>>> the way, at least one NCAR researcher was highly complementary of
>>> the Parco paper and said that it really clarified Swift for her).
>>> 
>>> - Mike
>>> 
>>> ----- Original Message -----
>>>>> -- The simplicity of script-code seems deceptive!?!
>>>> 
>>>> If that's refering to the script in the lower left corner, I guess
>>>> it
>>>> omits a bunch of stuff. Whether that's "deceptive" is rather
>>>> subjective -
>>>> for example, its ommitting the mappings (which at one point were a
>>>> big
>>>> part of the swift story). Now its probably ok to cough politely and
>>>> get
>>>> away with that for str_roots, but the code as given seems to assign
>>>> data_files[0] many times without it being clear from the fragment
>>>> that
>>>> data_files[0] is going to be different each time round (presumably
>>>> it
>>>> is?). Mapping often seems to be a user-effort-sink, so discarding
>>>> them
>>>> and
>>>> then saying "easy!" is kinda cheeky...
>>>> 
>>>> _______________________________________________
>>>> Swift-devel mailing list
>>>> Swift-devel at ci.uchicago.edu
>>>> http://mail.ci.uchicago.edu/mailman/listinfo/swift-devel
>>> 
>>> --
>>> Michael Wilde
>>> Computation Institute, University of Chicago
>>> Mathematics and Computer Science Division
>>> Argonne National Laboratory
>>> 
>>> _______________________________________________
>>> Swift-devel mailing list
>>> Swift-devel at ci.uchicago.edu
>>> http://mail.ci.uchicago.edu/mailman/listinfo/swift-devel
>> 
>> --
>> Daniel S. Katz
>> University of Chicago
>> (773) 834-7186 (voice)
>> (773) 834-3700 (fax)
>> d.katz at ieee.org or dsk at ci.uchicago.edu
>> http://www.ci.uchicago.edu/~dsk/
> 
> -- 
> Michael Wilde
> Computation Institute, University of Chicago
> Mathematics and Computer Science Division
> Argonne National Laboratory
> 

-- 
Daniel S. Katz
University of Chicago
(773) 834-7186 (voice)
(773) 834-3700 (fax)
d.katz at ieee.org or dsk at ci.uchicago.edu
http://www.ci.uchicago.edu/~dsk/