From dwells Sun Dec 6 16:46:26 1992 Status: RO X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil] ["1257" "Sun" "6" "December" "92" "16:46:25" "EST" "Don Wells" "dwells " nil "26" "Is anyone not ready to vote?" "^From:" nil nil "12"]) Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AA10546; Sun, 6 Dec 92 16:46:26 EST Return-Path: Message-Id: <9212062146.AA10540@fits.cv.nrao.edu> From: dwells (Don Wells) Sender: wfc-request@fits.CV.NRAO.EDU To: wfc Subject: Is anyone not ready to vote? Date: Sun, 6 Dec 92 16:46:25 EST Dear WFC, I would like to call for the formal WFC vote on the three proposals on Wednesday 09 December, two days from now. Does anyone object to this? Does anyone have a question about one of the proposals? Please reply to "wfc@nrao.edu" or to me privately if you are not ready to vote or have a question. I don't want to rush anyone, but if everyone is ready there is no reason for further delay. Regarding the voting procedure, my intention is to ask each of you to reply to me by Email specifying YES/NO/ABSTAIN on each of the three proposals, and when I have received all of the replies, I will post a listing of who voted in each category on each proposal. The three proposals are in the directory FITS/Documents on anonymous FTP server fits.cv.nrao.edu [192.33.115.8]: -r--r--r-- 1 dwells 2183 Oct 1 1991 FITS_blocking90.txt -r--r--r-- 1 dwells 179394 Sep 23 1991 X_bintable.ps -r--r--r-- 1 dwells 94980 Jun 4 1992 X_image.ps -Don Donald C. Wells Associate Scientist dwells@nrao.edu National Radio Astronomy Observatory +1-804-296-0277 520 Edgemont Road Fax= +1-804-296-0278 Charlottesville, Virginia 22903-2475 USA 78:31.1W, 38:02.2N From MAILER-DAEMON Sun Dec 6 17:54:53 1992 Status: RO X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil] ["1356" "Sun" "6" "December" "92" "17:52:14" "EST" "Mail Delivery Subsystem" "MAILER-DAEMON " nil "35" "Returned mail: unknown mailer error 1" "^From:" nil nil "12"]) Return-Path: Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AB10598; Sun, 6 Dec 92 17:52:14 EST Message-Id: <9212062252.AB10598@fits.cv.nrao.edu> From: MAILER-DAEMON (Mail Delivery Subsystem) To: wfc-request Subject: Returned mail: unknown mailer error 1 Date: Sun, 6 Dec 92 17:52:14 EST ----- Transcript of session follows ----- 421 noao.edu: Host noao.edu is down, will keep trying for 1 week /bin/sh: /fits/dwells/FITS/wfc.log: Permission denied 554 "|/bin/sh -c 'cat >> /fits/dwells/FITS/wfc.log'"... unknown mailer error 1 ----- Unsent message follows ----- Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AA10596; Sun, 6 Dec 92 17:52:14 EST Return-Path: Message-Id: <9212062252.AA29494@helios.UCSC.EDU> To: dwells@fits.CV.NRAO.EDU (Don Wells) Cc: wfc@fits.CV.NRAO.EDU, kibrick@lick.UCSC.EDU Subject: Re: Is anyone not ready to vote? In-Reply-To: Your message of "Sun, 06 Dec 92 16:46:25 EST." <9212062146.AA10540@fits.cv.nrao.edu> Date: Sun, 06 Dec 92 14:52:40 -0800 From: "Bob Kibrick -- kibrick@helios.ucsc.edu" Sender: wfc-request@fits.CV.NRAO.EDU Don, I will have difficulty reviewing any of the material by December 9, as I am home recovering from abdominal surgery which I underwent on December 3. Also, it was not until December 2 that I received notification that I was on the committee. However, I do not want to hold up the rest of the committee if there is urgency in getting the current vote completed. If that is the case, then I will most likely respectfully ABSTAIN from the current round of voting. Regards, Bob Kibrick UCO/Lick Observatory From MAILER-DAEMON Mon Dec 7 09:52:03 1992 Status: RO X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil] ["933" "Mon" "7" "December" "92" "09:50:35" "EST" "Mail Delivery Subsystem" "MAILER-DAEMON " nil "25" "Returned mail: unknown mailer error 1" "^From:" nil nil "12"]) Return-Path: Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AB12846; Mon, 7 Dec 92 09:50:35 EST Message-Id: <9212071450.AB12846@fits.cv.nrao.edu> From: MAILER-DAEMON (Mail Delivery Subsystem) To: wfc-request Subject: Returned mail: unknown mailer error 1 Date: Mon, 7 Dec 92 09:50:35 EST ----- Transcript of session follows ----- /bin/sh: /fits/dwells/FITS/wfc.log: Permission denied 554 "|/bin/sh -c 'cat >> /fits/dwells/FITS/wfc.log'"... unknown mailer error 1 ----- Unsent message follows ----- Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AA12844; Mon, 7 Dec 92 09:50:35 EST Return-Path: Date: Mon, 7 Dec 1992 09:50 EST From: HENDEN@MPS.OHIO-STATE.EDU Subject: votes To: wfc@NRAO.EDU Message-Id: <01GS13CLK3LS8WW8SV@MPS.OHIO-STATE.EDU> X-Envelope-To: wfc@nrao.EDU X-Vms-To: IN%"wfc@nrao.edu" Sender: wfc-request@fits.CV.NRAO.EDU Don, I don't want to muddy the water, or delay the voting, but there has been some discussion on the net regarding some of these proposals. If someone has these discussions archived, I think it would be wise to post a summary or the actual messages on fits.cv.nrao.edu so that the voters have a more complete pro/con list before voting. Arne From MAILER-DAEMON Mon Dec 7 10:04:16 1992 Status: RO X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil] ["666" "Mon" "7" "December" "92" "10:03:18" "EST" "Mail Delivery Subsystem" "MAILER-DAEMON " nil "17" "Returned mail: unknown mailer error 1" "^From:" nil nil "12"]) Return-Path: Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AB12884; Mon, 7 Dec 92 10:03:18 EST Message-Id: <9212071503.AB12884@fits.cv.nrao.edu> From: MAILER-DAEMON (Mail Delivery Subsystem) To: wfc-request Subject: Returned mail: unknown mailer error 1 Date: Mon, 7 Dec 92 10:03:18 EST ----- Transcript of session follows ----- /bin/sh: /fits/dwells/FITS/wfc.log: Permission denied 554 "|/bin/sh -c 'cat >> /fits/dwells/FITS/wfc.log'"... unknown mailer error 1 ----- Unsent message follows ----- Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AA12882; Mon, 7 Dec 92 10:03:18 EST Return-Path: Date: Mon, 7 Dec 92 08:03:08 MST From: tody@noao.edu (Doug Tody) Message-Id: <9212071503.AA05918@lepus.tuc.noao.edu> To: wfc@NRAO.EDU Subject: FITS proposals / discussions Sender: wfc-request@fits.CV.NRAO.EDU Arne, the sci.astro.fits discussions appear to already be archived on fits.nrao.edu, in FITS/sci_astro_fits. - Doug From MAILER-DAEMON Mon Dec 7 12:50:15 1992 Status: RO X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil] ["3048" "Mon" "7" "December" "92" "12:46:53" "EST" "Mail Delivery Subsystem" "MAILER-DAEMON " nil "75" "Returned mail: unknown mailer error 1" "^From:" nil nil "12"]) Return-Path: Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AB12977; Mon, 7 Dec 92 12:46:53 EST Message-Id: <9212071746.AB12977@fits.cv.nrao.edu> From: MAILER-DAEMON (Mail Delivery Subsystem) To: wfc-request Subject: Returned mail: unknown mailer error 1 Date: Mon, 7 Dec 92 12:46:53 EST ----- Transcript of session follows ----- 421 nssdca.gsfc.nasa.gov: Host nssdca.gsfc.nasa.gov is down, will keep trying for 1 week 421 ndadsa.gsfc.nasa.gov: Host ndadsa.gsfc.nasa.gov is down, will keep trying for 1 week /bin/sh: /fits/dwells/FITS/wfc.log: Permission denied 554 "|/bin/sh -c 'cat >> /fits/dwells/FITS/wfc.log'"... unknown mailer error 1 ----- Unsent message follows ----- Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AA12975; Mon, 7 Dec 92 12:46:53 EST Return-Path: Date: Mon, 7 Dec 92 12:47:13 EST From: pence@tetra.gsfc.nasa.gov (William Pence) Message-Id: <9212071747.AA02715@tetra.gsfc.nasa.gov> To: wfc@fits.CV.NRAO.EDU Subject: Re: Is anyone not ready to vote? Cc: pence@tetra.gsfc.nasa.gov Sender: wfc-request@fits.CV.NRAO.EDU As a point of clarification, can someone please verify that this vote only formally applies to the main body of each FITS proposal, and not to the appendices. In the case of the binary table extension, I would like to see a small change made to Appendix B (see below), and I would also like to see a paragraph added somewhere (perhaps to Appendix A, or possibly as a separate appendix) describing the convention for defining arrays of fixed-length character strings via the TFORMn = 'rAw' syntax. And, assuming that this vote does not cover the appendices, just what is the procedure for revising the appendices or submitting new ones. -Bill Pence ----------------------------------------------------------------------- Proposed change to Appendix B of the 'Binary Table Extension to FITS' The FITSIO software package supports an extension to the format of the TFORM keyword in binary tables to handle arrays of fixed length characters strings. The supported format has the form: TFORMn = 'rAw' where 'r' is an integer specifying the total number of characters in the field, and 'w' is an optional integer specifying the width of a unit character string within the field. For instance, '20A10' indicates that the field is 20 characters in length and contains 2 10-character strings. Note that this format for the TFORM keyword is legal because the binary table definition document specifically allows any additional characters to follow the datatype code character. Appendix B of the binary table definition document proposes that the TFORM keyword may have the form: rPt(maxelem) where t is a character denoting the datatype of the array data (L, X, B, I, J, A, etc.). I would like to see this modified slightly to have the form: rPtw(maxelem) where the 'w' is an optional integer used to specify the length of a unit string when the field contains an array of characters (i.e., then 't' = A). An example of this useage is: TFORMn = 'PA8(80)' which indicates that the field contains a varable number of 8-character strings, and that the field has a maximum width of 80 characters (= 10 strings). This proposed convention is optional and does not preclude other conventions (e.g., the TDIM convention described in Appendix A). From MAILER-DAEMON Mon Dec 7 16:14:00 1992 Status: RO X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil] ["3142" "Mon" "7" "December" "92" "15:42:08" "EST" "Mail Delivery Subsystem" "MAILER-DAEMON " nil "56" "Returned mail: unknown mailer error 1" "^From:" nil nil "12"]) Return-Path: Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AB13036; Mon, 7 Dec 92 15:42:08 EST Message-Id: <9212072042.AB13036@fits.cv.nrao.edu> From: MAILER-DAEMON (Mail Delivery Subsystem) To: wfc-request Subject: Returned mail: unknown mailer error 1 Date: Mon, 7 Dec 92 15:42:08 EST ----- Transcript of session follows ----- 451 net hang reading from hapuna.keck.hawaii.edu: Connection timed out during RCPT wait with hapuna.keck.hawaii.edu /bin/sh: /fits/dwells/FITS/wfc.log: Permission denied 554 "|/bin/sh -c 'cat >> /fits/dwells/FITS/wfc.log'"... unknown mailer error 1 ----- Unsent message follows ----- Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AA13034; Mon, 7 Dec 92 15:42:08 EST Return-Path: Date: Mon, 7 Dec 92 13:41:55 MST From: tody@noao.edu (Doug Tody) Message-Id: <9212072041.AA07209@lepus.tuc.noao.edu> To: wfc@NRAO.EDU Subject: binary tables discussion Sender: wfc-request@fits.CV.NRAO.EDU Bill - I cannot answer your question on what our vote formally means. My expectation is that we are voting only on the proposals as they currently stand. In the case of the BT proposal, my understanding is that technically the vote only refers to the main proposal, however the intention is to include the current appendices in the final document. Regarding your specific comments on arrays of strings - I think this should be the subject of either a layered convention or an additional appendix. It does not appear necessary to modify the core BT proposal to add support for this type of data. The important question is whether it is worth while holding up the proposal until this issue is resolved. If you do not feel that the proposal should be approved until this issue has been resolved, probably you should vote no on the proposal, or try to convince the committee to delay until the matter is resolved. If you feel that the matter can be dealt with as an addition to the current proposal and there is insufficient reason to delay adoption of the current proposal, then you should vote yes. A possible compromise might be to approve the proposal as it currently stands but with a note that important questions remain regarding how to handle character data (as well as array data). Regarding your specific proposal for arrays of strings - I agree that the current proposal does not provide much guidance regarding how to deal with this type of data. However, I cannot agree with the convention you propose. I do not like arrays of fixed length blank filled character strings. This should not be the only approach for representing such data suggested by the BT proposal. I could expand upon this but I don't think it is appropriate to discuss the matter in detail here. It is sufficient to note that the matter is still controversial and we do not know yet how best to represent either complex string data or multidimensional array data in binary tables. I could draft a layered convention for the representation of arrays of strings combining, e.g., arrays of fixed length strings and sequences of delimited substrings as alternate forms within the same convention, however I think there might be complications with this and attemping to get it in the current proposal would delay things. Since it can be added later my preference would be to proceed with this separately. Multidimensional array support is related and should be handled the same way. - Doug From MAILER-DAEMON Tue Dec 8 15:38:16 1992 Status: RO X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil] ["5556" "Tue" "8" "December" "92" "15:36:40" "EST" "Mail Delivery Subsystem" "MAILER-DAEMON " nil "108" "Returned mail: unknown mailer error 1" "^From:" nil nil "12"]) Return-Path: Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AB14359; Tue, 8 Dec 92 15:36:40 EST Message-Id: <9212082036.AB14359@fits.cv.nrao.edu> From: MAILER-DAEMON (Mail Delivery Subsystem) To: wfc-request Subject: Returned mail: unknown mailer error 1 Date: Tue, 8 Dec 92 15:36:40 EST ----- Transcript of session follows ----- /bin/sh: /fits/dwells/FITS/wfc.log: Permission denied 554 "|/bin/sh -c 'cat >> /fits/dwells/FITS/wfc.log'"... unknown mailer error 1 ----- Unsent message follows ----- Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AA14357; Tue, 8 Dec 92 15:36:40 EST Return-Path: Date: Tue, 8 Dec 92 15:36:59 EST From: pence@tetra.gsfc.nasa.gov (William Pence) Message-Id: <9212082036.AA03827@tetra.gsfc.nasa.gov> To: wfc@NRAO.EDU Subject: Re: binary tables discussion Sender: wfc-request@fits.CV.NRAO.EDU In reply to Doug Tody's message on 7 Dec 1992: On whether to vote now, or postpone it until the technical issues are resolved: My preference would be to hold off on the vote until this character string issue is resolved. I think this is an important issue; I also think we are actually close to an agreement on how to support both fixed and delimited strings (as has been discussed previously in the sci.astro.fits group) so I would think we could have a new draft in a week or 2 at the most. The changes would only affect the appendices and not the main body of the binary table proposal. I have lobbyed for this change for a long time to whoever listens to astro.sci.fits, but now that the WFC has been formed I would like to see this issue seriously addressed. On the other hand, if our Chairman feels that there is a need to have a vote on the proposals immediately, then I would go along with the decision. But I seems to me that a couple week's delay would not hurt much. Regarging the vote, I still don't really understand what it means to vote on a document that states (in Appendix B) that "this facility is undergoing trials and is not currently part of the main binary table definition". It seems to me we need to either approve or disapprove this convention, and remove the wafflely words about it being under trial. The HEASARC, for one, is already using this convention is some of it's FITS files, so we would like to see it approved as more than just a 'trial'). On some of the technical issues: 1. Defining the string length in a FITS table is of fundamental importance and is distinctly different from defining multidimensional arrays. In order to read or write to a character ('A') column in a binary table it is *required* that you know how to determine the character string length since users are almost always interested in reading or writing strings, not individual characters. As an example, suppose a user asks me to write an array of character strings to a character column which has TFORM = '40A'. I am unable to do this however until you also tell me how to pack the strings into the column; should I write one string into each 40 character field, or should I perhaps write 4 strings, each with a maximum of 10 characters into the same field. Or, should I pack as many strings as possible into the field using a delimiter code character?? The point is that defining the string length (regardless of whether it is fixed length or variable length) is of such fundamental importance in order to correctly interpret a character column that this information should be directly encoded into the TFORM string, and not relegated to an optional TDIM keyword which also serves a different purpose. To just briefly follow up on this arguement, note that this problem does not affect any other datatype column. For instance, if I am asked to write 50 real*4 numbers into a column that has TFORM = '20E' then there is little ambiguity about what to do: simply write 20 numbers into each of the first 2 rows, and the remaining 10 numbers into the beginning of the 3rd row. (Or one could write one number into the first element of the first 50 rows. But this ambiguity can easily be resolved by defining a convention within any given user interface). At this level of data access there is never any need to know anything about the higher-level dimensionality of the data that is being written or read as defined by the TDIM convention. 2. On the issue of fixed length vs. delimited character strings. My proposal is quite conservative in that it is only suggesting a much more useful way of specifying the size of fixed-length strings instead of using the TDIM convention described in Appendix A. (Note that the TDIM convention also doesn't provide any way to define delimited string fields). To raise the possibility of allowing arrays of variable length strings (i.e., delimited strings) in character columns is much more controversial. But in any case, my proposal for specifying fixed length strings in no way prohibits or limits defining additional conventions for variable length strings. either now, or in the future. I have thought quite a bit about how one might support arrays of delimited strings in FITS files through the FITSIO interface, but it raises a number of thorny issues, and it doesn't seem to provide much practical benefit when dealing with fixed format structures like FITS files. The concept of delimited strings may be great in an environment where you have a dynamic memory allocation capablility, but in FITS files with fixed length fields and rows, having delimited strings is not really very useful; for instance you don't gain any savings in file size, since each column has a fixed width (except for variable length arrays, which is not at issue here). So it seems to me that allowing for delimited-string arrays in character columns introduces a lot more problems than it is worth. -Bill Pence From MAILER-DAEMON Tue Dec 8 16:22:06 1992 Status: RO X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil] ["1843" "Tue" "8" "December" "92" "16:20:36" "EST" "Mail Delivery Subsystem" "MAILER-DAEMON " nil "36" "Returned mail: unknown mailer error 1" "^From:" nil nil "12"]) Return-Path: Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AB14398; Tue, 8 Dec 92 16:20:36 EST Message-Id: <9212082120.AB14398@fits.cv.nrao.edu> From: MAILER-DAEMON (Mail Delivery Subsystem) To: wfc-request Subject: Returned mail: unknown mailer error 1 Date: Tue, 8 Dec 92 16:20:36 EST ----- Transcript of session follows ----- /bin/sh: /fits/dwells/FITS/wfc.log: Permission denied 554 "|/bin/sh -c 'cat >> /fits/dwells/FITS/wfc.log'"... unknown mailer error 1 ----- Unsent message follows ----- Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AA14396; Tue, 8 Dec 92 16:20:36 EST Return-Path: Date: Tue, 8 Dec 92 14:20:24 MST From: tody@noao.edu (Doug Tody) Message-Id: <9212082120.AA11270@lepus.tuc.noao.edu> To: wfc@NRAO.EDU Subject: binary tables discussion Sender: wfc-request@fits.CV.NRAO.EDU Bill, I agree that the issue of strings in binary tables is an important one. Although this has been discussed off and on all year, so far we haven't done anything about it. Probably there has been enough discussion and some sort of draft appendix to the BT document is called for. However, I think you are being extremely optimistic in thinking that this could be in hand in one or two weeks. I think this would take at least 1-2 man-days (1-2 calendar weeks) for, say, 2 or 3 of us working together to carefully generate a draft, then perhaps 2 weeks for the WFC and IAU WG to discuss it internally. We might then want to put the proposed appendix out on sci.astro.fits for public review for a time. I am willing to help with this, in fact I think I need to be involved, but my time is fully committed until after the AAS meeting January 3-7. Hence we cannot start this process until at least January and realistically it will take several months to complete. It remains possible that when we review the technical issues again carefully there will still be disagreement over whether this should be included in the current proposal. Hence there is no guarantee that an agreement would be reached, although I am confident that we could at least agree on a layered convention. - Doug From MAILER-DAEMON Tue Dec 8 17:11:59 1992 Status: RO X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil] ["1064" "Tue" "8" "December" "92" "17:10:43" "EST" "Mail Delivery Subsystem" "MAILER-DAEMON " nil "25" "Returned mail: unknown mailer error 1" "^From:" nil nil "12"]) Return-Path: Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AB14483; Tue, 8 Dec 92 17:10:43 EST Message-Id: <9212082210.AB14483@fits.cv.nrao.edu> From: MAILER-DAEMON (Mail Delivery Subsystem) To: wfc-request Subject: Returned mail: unknown mailer error 1 Date: Tue, 8 Dec 92 17:10:43 EST ----- Transcript of session follows ----- /bin/sh: /fits/dwells/FITS/wfc.log: Permission denied 554 "|/bin/sh -c 'cat >> /fits/dwells/FITS/wfc.log'"... unknown mailer error 1 ----- Unsent message follows ----- Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AA14481; Tue, 8 Dec 92 17:10:43 EST Return-Path: Date: Tue, 8 Dec 92 17:11:12 EST From: pence@tetra.gsfc.nasa.gov (William Pence) Message-Id: <9212082211.AA03892@tetra.gsfc.nasa.gov> To: wfc@NRAO.EDU Subject: Re: binary tables discussion Sender: wfc-request@fits.CV.NRAO.EDU If Doug's time estimates are correct, then I leave it to the Chairman to decide when to call for a vote. I would vote in favor of all the proposals as they stand, on the understanding that the appendices to the binary table proposal would be reviewed as soon as the appropriate people are available to do so. Should we appoint a subcommittee of WFC members to review and (possibly) redraft the appendices and resubmit them to the full WFC (and sci.astro.fits) for consideration? -Bill From MAILER-DAEMON Tue Dec 8 17:44:32 1992 Status: RO X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil] ["1661" "Tue" "8" "December" "92" "17:43:09" "EST" "Mail Delivery Subsystem" "MAILER-DAEMON " nil "32" "Returned mail: unknown mailer error 1" "^From:" nil nil "12"]) Return-Path: Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AB14544; Tue, 8 Dec 92 17:43:09 EST Message-Id: <9212082243.AB14544@fits.cv.nrao.edu> From: MAILER-DAEMON (Mail Delivery Subsystem) To: wfc-request Subject: Returned mail: unknown mailer error 1 Date: Tue, 8 Dec 92 17:43:09 EST ----- Transcript of session follows ----- /bin/sh: /fits/dwells/FITS/wfc.log: Permission denied 554 "|/bin/sh -c 'cat >> /fits/dwells/FITS/wfc.log'"... unknown mailer error 1 ----- Unsent message follows ----- Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AA14542; Tue, 8 Dec 92 17:43:09 EST Return-Path: Date: Tue, 8 Dec 92 15:42:57 MST From: tody@noao.edu (Doug Tody) Message-Id: <9212082242.AA11570@lepus.tuc.noao.edu> To: wfc@NRAO.EDU Subject: binary tables discussion Sender: wfc-request@fits.CV.NRAO.EDU > people are available to do so. Should we appoint a subcommittee of > WFC members to review and (possibly) redraft the appendices and > resubmit them to the full WFC (and sci.astro.fits) for consideration? Bill, in this case I think you and I should try to draft something we can agree on. If it is to be included in the BT document in the near future then the authors (currently Bill Cotton and myself) would have to agree to include it. Hence, the three of us form a logical group to thrash this out. I suggest we try that approach first. If later we decide to not include this as an appendix to the BT paper then an arbitrary set of people could be assembled and the result would be a separate paper describing a BT layered convention. Or, we could just agree to freeze the BT paper in its current state and proceed with the vote, then introduce this as a separate layered convention later. Either way I am interested in helping define this once time permits. Ultimately I imagine there will be many layered conventions for binary tables defining how to represent different types of data. - Doug From wfc-request@fits.CV.NRAO.EDU Wed Dec 9 12:12:33 1992 Status: RO X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil] ["3305" "Wed" "09" "December" "92" "16:00:47" "+0100" "pgrosbol@eso.org" "pgrosbol@eso.org" nil "60" "Re: binary tables discussion" "^From:" nil nil "12"]) Received: from cv3.cv.nrao.edu by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AA16148; Wed, 9 Dec 92 12:12:32 EST Received: from fits.cv.nrao.edu by cv3.cv.nrao.edu (4.1/DDN-DLB/1.13) id AA17419; Wed, 9 Dec 92 12:12:27 EST Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AA16145; Wed, 9 Dec 92 12:12:16 EST Return-Path: Message-Id: <9212091500.AA10678@ns2.hq.eso.org> From: pgrosbol@eso.org Sender: wfc-request@fits.CV.NRAO.EDU To: wfc@fits.CV.NRAO.EDU Subject: Re: binary tables discussion Date: Wed, 09 Dec 92 16:00:47 +0100 Garching, Dec. 9, 1992 Dear WGAS FITS Committee Members, I have followed the discussion in the WGAS FITS Committee concerning the BINTABLE proposal and think it would be useful to state some procedural points. The general procedure for accepting a proposal as a FITS standard is summarized in the following points: 1) General discussion to identify if a proposal is of general interest and reasonable to promote as a standard, 2) Submission to the regional FITS groups for detailed discussion, 3) Test implementation to verify that proposal is feasible and well defined, 4) Motions in support of the proposal in the regional FITS groups, 5) If all groups support the proposal, it is submitted to the IAU FITS WG for final review, 6) Full test and demonstration of transport of data between different sites and systems, 7) Final vote in the IAU FITS WG, 8) Proposal by IAU FITS WG to IAU General Assembly to recommend the proposal as a standard. The three new FITS proposals are currently at point 4 where the European FITS Committee already this summer has passed a motion in support of them. In case the three regional FITS groups (WGAS, Japanize, European) cannot reach agreement on a proposal, it would not be passed on to the IAU FITS WG. To resolve the current problem with character string in BINTABLE's I suggest that the WGAS FITS Committee try to find its own consensus on the matter. If that consensus is different from the proposal, the Chairman of WFC should inform the Chairman of the IAU FITS WG. He would then report it to the Chairmen of the other regional FITS groups with a request to consider supporting the new proposal with motions. The discussion under point 1 (through sci.astro.fits) should normally avoid that an official iteration of proposals is made between the different FITS groups. Although there was a discussion on the character string issue in the spring it was obviously not detailed enough. To avoid delays, I will already now ask the other FITS groups whether or not they would have any objections in including arrays of strings by allowing TFORM = 'wAn' formats. The issue was discussed in during the European FITS Committee meeting in May 1992. It was not considered so important that the proposal should be rejected and a revision made. The main argument was that one later could add a convention which would interpret 'wA' formats as 'wA1' and allow 'wAn' formats. Files written with 'wA' would then still conform. Personaly, I would support arrays of strings with the 'wAn' format. It would give a more symmetric definition compared to ASCII tables which include strings as a basic type. Further, it would solve some problems if we later would like to save definitions of multi-dimensional array in single columns e.g. corresponding to CUNIT and CTYPE. I believe it is reasonable to have this issue resolved before a vote is made and suggest that the WFC try to reach a consensus possibly through a revision of the proposal. Although the IAU FITS WG may make clarifications and minor revisions of a proposal during its review, it should not change basic issues like data types or formats allowed. Best regards, Preben Grosbol, ESO Chair, IAU FITS WG, Chair, European FITS Committee. From wfc-request@fits.CV.NRAO.EDU Tue Dec 15 19:00:30 1992 Status: RO X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil] ["1076" "Tue" "15" "December" "92" "15:59:40" "PST" "bob@ipac.caltech.edu" "bob@ipac.caltech.edu" nil "42" "Binary Table Character Strings 2 Cents" "^From:" nil nil "12"]) Received: from cv3.cv.nrao.edu by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AA27937; Tue, 15 Dec 92 19:00:29 EST Received: from fits.cv.nrao.edu by cv3.cv.nrao.edu (4.1/DDN-DLB/1.13) id AA04112; Tue, 15 Dec 92 19:00:24 EST Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AA27934; Tue, 15 Dec 92 18:59:51 EST Return-Path: Message-Id: <9212152359.AA04107@shatner.ipac.caltech.edu> From: bob@ipac.caltech.edu Sender: wfc-request@fits.CV.NRAO.EDU To: wfc@NRAO.EDU Subject: Binary Table Character Strings 2 Cents Date: Tue, 15 Dec 92 15:59:40 PST On the handling of character strings in FITS binary tables, here are the questions I see: 1. How should FITS represent "b bytes composed of n strings of w characters each" (b = n * w) ? (A) "TFORM = nAw" (B) "TFORM = bAw" (C) "TFORM = bA" and "TDIM = '(w,n)'" 2. What should "TFORM = wA" mean? (A) w characters of unspecified structure (B) one string of w characters (C) w strings of one character each (any or all may be nulls) 3. Should strings be allowed to be null terminated ? (A) yes, remaining bytes in string length are ignored (B) no, must blank fill The most logical (best) to me is "A" for each of the above. However, we have no existing binary table files at this time, and I suspect this is not the case for all institutions. There are probably existing archives of FITS binary table files that are inconsistent with the "A" choices above. If so, I'd be willing to support any of the other choices. Is there a way to get a list of any such archives ? Bob Narron From wfc-request@fits.CV.NRAO.EDU Wed Dec 16 11:40:50 1992 Status: RO X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil] ["2366" "Wed" "16" "December" "92" "11:40:55" "EST" "William Pence" "pence@tetra.gsfc.nasa.gov " nil "62" "Re: Binary Table Character Strings 2 Cents" "^From:" nil nil "12"]) Received: from cv3.cv.nrao.edu by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AA29206; Wed, 16 Dec 92 11:40:49 EST Received: from fits.cv.nrao.edu by cv3.cv.nrao.edu (4.1/DDN-DLB/1.13) id AA17667; Wed, 16 Dec 92 11:40:44 EST Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AA29203; Wed, 16 Dec 92 11:40:19 EST Return-Path: Message-Id: <9212161640.AA05066@tetra.gsfc.nasa.gov> From: pence@tetra.gsfc.nasa.gov (William Pence) Sender: wfc-request@fits.CV.NRAO.EDU To: wfc@NRAO.EDU Cc: pence@tetra.gsfc.nasa.gov Subject: Re: Binary Table Character Strings 2 Cents Date: Wed, 16 Dec 92 11:40:55 EST Bob Narron wrote: > On the handling of character strings in FITS binary tables, > here are the questions I see: > > > 1. How should FITS represent "b bytes composed of n strings > of w characters each" (b = n * w) ? > > (A) "TFORM = nAw" > > (B) "TFORM = bAw" > > (C) "TFORM = bA" and "TDIM = '(w,n)'" My FITSIO software was originally written to support (A). However, when it appeared that the European FITS committee had voted to approve the binary table definition document in its current form, then I modified FITSIO to only support the (B) format which is in compliance with the binary table definition document. This change occurred with version 3.21 of FITSIO which was released in July 1992. Note that under this convention I interpret TFORM = 'nA' to mean the same thing as TFORM = 'nAn' > 2. What should "TFORM = wA" mean? > > (A) w characters of unspecified structure > > (B) one string of w characters > > (C) w strings of one character each (any or all may be nulls) In the absence of any other information or conventions, I think this should mean that the field contains one string, which is either blank filled or null terminated. > 3. Should strings be allowed to be null terminated ? > > (A) yes, remaining bytes in string length are ignored > > (B) no, must blank fill The current draft of the binary table definition document states that "A character string may be terminatated before its explicit length by an ASCII NULL character". I see no reason to change this. > The most logical (best) to me is "A" for each of the above. > > However, we have no existing binary table files at this time, > and I suspect this is not the case for all institutions. > There are probably existing archives of FITS binary table > files that are inconsistent with the "A" choices above. > If so, I'd be willing to support any of the other choices. > Is there a way to get a list of any such archives ? I'm not aware of any archival FITS data which has used anything other than the simple TFORM = 'nA' for ASCII character columns in binary tables. I think the main objection that has been raised to this convention is that it is not general enough and does not define how to specify arrays of terminated strings. This is what we still need to work on. -Bill Pence From wfc-request@fits.CV.NRAO.EDU Wed Dec 23 12:52:44 1992 Status: RO X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil] ["223" "Wed" "23" "December" "1992" "12:51:59" "-0500" "teuben@astro.UMD.EDU" "teuben@astro.UMD.EDU" nil "7" "FITS images+colorLUT's" "^From:" nil nil "12"]) Received: from cv3.cv.nrao.edu by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AA04313; Wed, 23 Dec 92 12:52:43 EST Received: from fits.cv.nrao.edu by cv3.cv.nrao.edu (4.1/DDN-DLB/1.13) id AA25134; Wed, 23 Dec 92 12:52:36 EST Received: by fits.cv.nrao.edu (4.1/DDN-DLB/1.5) id AA04310; Wed, 23 Dec 92 12:52:28 EST Return-Path: Message-Id: <199212231751.AA05338@saturn.astro.umd.edu> From: teuben@astro.UMD.EDU Sender: wfc-request@fits.CV.NRAO.EDU To: wfc@NRAO.EDU Subject: FITS images+colorLUT's Date: Wed, 23 Dec 1992 12:51:59 -0500 Has anybody thought about a standard extension which would describe a color lookup table, which could be appended to images? Also needed is a description of the color model (RGB/HSV etc.) and other details. Peter Teuben