Date: Sat, 8 Dec 2001 17:52:51 -1000 From: Doug Summers To: "'denault@jeans.ifa.hawaii.edu'" Subject: RE: version 1.1 of the LTCS URL interface spec Parts/attachments: 1 Shown 114 lines Text 2 71 KB Application ---------------------------------------- Tony, Thanks for pointing the URL spec errors out....nobody else bothered to tell me! I've changed the doc accordingly; here's the latest updates (I removed the exponential formats as well....nobody will miss them!). I'll test the IRTF URL tonight. As for the stale data, I've spent some time thinking about this and there may be a way to handle this....I might just have the collector tell the GA that the laser_impacted flag is false under the stale scenario (and include a control flag so I can figure out that it was the collector who initiated the state change; I'll then log that this occurred due to stale data). Don't know that I can implement it immediately, but I will work it. Hope you have a great weekend...I think I'll take Sunday off and enjoy the beach. Doug -----Original Message----- From: Tony Denault [mailto:denault@jeans.ifa.hawaii.edu] Sent: Friday, December 07, 2001 5:56 PM To: Doug Summers Subject: RE: version 1.1 of the LTCS URL interface spec On Fri, 7 Dec 2001, Doug Summers wrote: > Hi Tony, > > The LTCS is partitioned into two distinct unix processes. One of these > collects the data from all telescopes on the mountain, and the other > processes the inputs through a geometric analysis (GA) model. The collector > will see if timestamps are stale and complain, but the geometric model won't > know about this. It continues to use whatever the last information it had. > Because of this, it's necessary to set the URL to be LASER_IMPACTED=NO at > shutdown time. Then it doesn't matter if it's read or not, the GA will know > that the positions don't matter. Otherwise, the collector just alarms. > > Do you think this will be a big problem for us? yes, I think if the data is stale ( XX minute old) it should be ignore. Maybe XX can be something big like 1 hour. > > > a. spelling problem with "TIMESTAMPE1, TIMESTAMPE2": Need to drop the E > It should be "TIMESTAMP1" and "TIMESTAMP2" Fixed...I have fat fingers and can't spell syndrome. > b. You are the only site to use exponential formats in the URL. I haven't > seen this or tested it yet (although I know it says in the spec we can > handle it). I'm not sure if you can convert to non-exponential format > or not, but if it's easy, you might consider it for RA/DEC & FOV. If > it's > a problem to do this, I'll test it. I just haven't yet and I'm worried > that the Gemini code (collector) may not handle it. If not, I'll need > to > go into their code and adjust, but I'd rather not do that before Tuesday > if I can help it. Is this something you can easily change in case it > doesn't work? I changed it to non-exponential, which I like better and used first. But your specification sample showed the exponential formats. So I though this was your preference. > c. While it is allowed to set the LOG_DATA flag to OFF, by agreement of the > MKLGSTWG members, this flag should be set to ON during the initial LTCS > software testing period. The effect of having it off is that I will > lose > all knowledge of where you are pointing, and we won't be able to debug > any problems that arise as a result of the laser crossing your guider. > If you can change this to ON for now, that will be great. > I will set it to ON. BTW: some suggestions for you specification document: 1. Page 2, your sample give TIMESTAMP as float (0.0) while the Keyword pages say it has long percision. Print the sample as a long. 2. Page 2, your same shows LOG_DATA = YES, the keyword says it is OFF / ON. 3. Also page 2, I would shoe EQUINOX in its float format ie: 2000.0 And most important...It friday and almost 6 pm. Time to go home. (I think I take this suggest myself) Hope you have a good weekend, Tony /-------------------------------------------------------------------------\ | Tony Denault | Email: denault@irtf.ifa.hawaii.edu | | NASA IRTF, Institute for Astronomy | Phone: (808) 933-0893 | | 640 North Aohoku Place | Fax: (808) 933-0737 | | Hilo, Hawaii 96720 | | \-------------------------------------------------------------------------/