From iaufwg-request Wed Apr 20 06:52:02 1994 Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AA26290; Wed, 20 Apr 94 06:52:02 EDT Return-Path: Message-Id: <9404201051.AA08295@ns2.hq.eso.org> To: iaufwg@fits.CV.NRAO.EDU Subject: Call for Vote on IMAGE and Blocking proposals Date: Wed, 20 Apr 94 12:51:53 +0200 From: pgrosbol@eso.org Sender: iaufwg-request@fits.CV.NRAO.EDU Garching, April 20, 1994 Dear IAU-FWG members, I have hesitated calling the final vote on the three proposal due to the on-going unsigned integer discussions. It mainly concerns the BINTABLE proposal and should no stall a final vote of the IMAGE and Blocking proposals. Thus, I hereby call for your vote on the IMAGE and Blocking proposals including the clarifications given below: 1) IMAGE extension The text of the proposal is available through anonymous ftp at fits.cv.nrao.edu in the directory /FITS/Documents as files: 94980 Jun 4 1992 X_image.ps 12973 Mar 12 1992 X_image.tex Clarification: Only PCOUNT=0/GCOUNT=1 is allowed (as specified in Table 1). Section 3.1.5 and 3.1.6 shall be interpreted in this sense. 2) Blocking proposal The text of the proposal is available through anonymous ftp at fits.cv.nrao.edu in the directory /FITS/Documents as files: 2183 Oct 1 1991 FITS_blocking90.txt Clarification: The sentence 'Reading an incomplete FITS logical record should be regarded as an end-of-file.' in section 2 is replaced by: 'After detection of an end-of-file on reading, any incomplete logical FITS record should be disregarded.' I hereby ask you to submit (by e-mail to pgrosbol@eso.org) your votes on each of the two proposals with a clear YES or NO (i.e. indicate if you vote YES/NO to 1) the IMAGE extension and if you vote YES/NO to 2) the Blocking proposal). The deadline for votes is: Friday, May 6, 1994, 18.00 GMT You are welcome to submit votes earlier. I will remind you again a few days before the deadline. For BINTABLE, I ask you for your opinion on two additional issues: a) The possible scaling of values pointed to by a P-format field was mentioned on the news-group. To make that possible through a conversion later the authors of the BINTABLE proposal suggest to change the sentence page 4 point 6: 'Note: TSCALnnn and TZEROnnn are not defined for A, L, P, or X format fields' to: 'Note: TSCALnnn and TZEROnnn are not defined for A, L, or X format fields. The anticipated meaning of TSCALnnn and TZEROnnn for P fields is described in Appendix A.' b) A more severe point is possible inclusion of unsigned integer types. It has been argued, on the net, that one should allow and define special format types for unsigned integers for BINTABLEs. Such a change would be significant and call for a re-approval of a modified BINTABLE proposal by the local FITS groups! Thus, two approaches can be made: b1) We can regard the suggestion for unsigned integer types not justified and go ahead with the formal approval of the BINTABLE proposal with the minor clarifications, or b2) Postpone a vote in the IAU FITS WG until the local FITS Committees have had time to consider and approve a revised version of the BINTABLE proposal. (This may well take more than one year.) I do not believe that unsigned integers are essential for a transport format like FITS and would therefore recommend to take option b1). Further, there seems not to be a consensus in the community on the introduction of unsigned integers. It is a rather important issue and I ask you to let me know your opinion especially if you want to postpone the vote (b2). If more that half of the IAU FITS WG express their support of option (b2) through e-mail to me before May 6, I will not call a vote on the BINTABLE before the local committees have been heard otherwise I will go ahead with a formal vote. Best regards, Preben Grosbol Chairman, IAU FITS WG From iaufwg-request Thu Apr 21 09:01:25 1994 Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AA28685; Thu, 21 Apr 94 09:01:25 EDT Return-Path: Date: Thu, 21 Apr 1994 9:02:38 -0400 (EDT) From: "BARRY M. SCHLESINGER" To: iaufwg@fits.CV.NRAO.EDU Message-Id: <940421090238.2ac00d17@NSSDCA.GSFC.NASA.GOV> Subject: BINTABLE and unsigned integers Sender: iaufwg-request@fits.CV.NRAO.EDU Dear IAU-FWG members, My first answer was made using REPLY and so apparently went to the REQUEST branch. So if it shows up later, now you'll know why. > For BINTABLE, I ask you for your opinion on two additional issues: > a) The possible scaling of values pointed to by a P-format field > was mentioned on the news-group. To make that possible through > a conversion later the authors of the BINTABLE proposal suggest > to change the sentence page 4 point 6: > 'Note: TSCALnnn and TZEROnnn are not defined for A, L, P, > or X format fields' > to: > 'Note: TSCALnnn and TZEROnnn are not defined for A, L, or > X format fields. The anticipated meaning of TSCALnnn and > TZEROnnn for P fields is described in Appendix A.' I suggested the approach of having TSCALn and TZEROn for P fields apply to the heap data designated by the array descriptor in response to a posting from Fermilab that there was no way to scale heap data. Bill Pence concurred that the inability to scale heap data was probably an oversight. This wording, along with appropriate revisions to Appendix A (which, I understand, is not a subject of the vote) should provide for scaling of heap data. > b) ... unsigned integer types. > ... Such a change would be significant and call for a re-approval > of a modified BINTABLE proposal by the local FITS groups! Thus, > two approaches can be made: > b1) We can regard the suggestion for unsigned integer types > not justified and go ahead with the formal approval of the > BINTABLE proposal with the minor clarifications, > or > b2) Postpone a vote in the IAU FITS WG until the local FITS > Committees have had time to consider and approve a revised > version of the BINTABLE proposal. (This may well take more > than one year.) I would propose an alternative b3), (or perhaps b1')). That approach would be to proceed now with approval of BINTABLE as it stands without prejudice as to the future addition of unsigned integers. The idea would be that since there is no community consensus at present in favor of unsigned integers, it would clearly be inappropriate to include them now, but, should the community so decide at a later time, they could then be added to BINTABLE as one or more additional data types. Approval of such an addition would follow the regular procedure. There is a precedent for adding new data types to an existing format: the Floating Point Agreement. If it were decided to incorporate any unsigned integer data types, they would be additions to BINTABLE as it is and not change the extension rules in such a way as to cause any violation of the rules by existing files. This approach allows us to formally approve the widely used and much needed BINTABLE now, at the same time not forcing an immediate decision on unsigned integers. Best regards, Barry Schlesinger