RLIN PUT SPECIFICATION/DEVELOPMENT WORK
Date: Thu, 28 Jul 1994 12:00:11 -1000
From: "K. Smith-Yoshimura"
To: macmill@uhunix.uhcc.Hawaii.Edu
Subject: More Follow-up
And meanwhile, Bruce Washburn has been communicating with Martha Chantiny, and
we should be able to get a trial account up and running for your technical
services staff to search RLIN for source copy and FTP records next week. More
on this soon.
Cheers and Aloha, Karen S-Y
Date: Wed, 27 Jul 1994 11:37:08 -1000
From: "Bruce Washburn"
To: chantiny@uhunix.uhcc.hawaii.edu
Subject: RLIN "PUT" command for FTP export of MARC records
I've attached some general information about the new RLIN "PUT" command,
below. My understanding is that a trial account is being set up for
your use during the month of August, and I'll let you know when this is
enabled for use of the PUT command.
Right now you can connect to one of RLG's FTP servers to get a sample
export file (include 6 MARC records) and a sample log file for those 6
export transactions. To get these files using your FTP client:
Connect to Lyra.Stanford.Edu
Logon as "anonymous"
Change to the directory "upload"
The file of six exported MARC records has the name
PUT.072794.EXP
The log file for these 6 transactions has the name
ulog940727
Date: Fri, 29 Jul 1994 09:06:30 -1000
From: "Bruce Washburn"
To: chantiny@uhunix.uhcc.hawaii.edu
Subject: PUT command test access
Dou should be receiving more information shortly on the RLIN trial account
that is being set up for your evaluation of the PUT command for RLIN MARC
record export, and for other RLIN technical processing services. When you
receive this RLIN account, you can use it to export records via the PUT
command to RLG's FTP server.
To test the PUT command, connect to the RLIN system (on a dedicated line,
via dialup, or over the Internet) and logon with your trial account.
Enter the command
cal rlin (cat)
to start your RLIN session, then begin searching. You can export any
single record that you have displayed during your search session. Just
enter the command
PUT
and the currently displayed record will be exported in MARC format to the
FTP server. Details on how the files are stored on the server, and how
they can be retrieved, are given below.
The PUT command can also include a parameter for including a string of local
information in the exported record. For example, if you find an RLIN record
to export and you'd like to associate that record with one in your local
system, you can include the local system record ID as a parameter on the
PUT command, e.g.,
PUT 856789
and the parameter value will appear in a 908 field in the exported record.
This feature would allow a local system to look for contents in a 908 field
to match exported RLIN records with existing local system records, to
make record transfer and overlay in the local system easier.
Details on the RLIN MARC record format are included in the document "The
RLIN MARC Record: Description and Interpretation", available from RLG.
Information on FTP server and PUT command setup:
With your RLIN trial account logged on under the library identifier HAUG:
Records will be exported to a file beginning with 'PUT', and
ending with '.EXP'
One file will be created per day, containing all exported records
The full file name will take the form
PUT.mmddyy.EXP
e.g.,
PUT.080594.EXP
Each day's export activity will also be reported in a separate
user log file, with the file name beginning with 'ulog' and
ending with the year, month, and day, e.g.,
ulog940805
These files will be stored in a directory on RLG's FTP server.
The FTP server address is
rlgftp.stanford.edu
- or -
36.26.0.15
To obtain HAUG account export files, logon to rlgftp.stanford.edu with
the user name
HAUG
and the password
AXXXXXXXXXX
RLG FTP server user names and passwords are case-sensitive, so must
be entered in upper-case to be recognized by the server. Stored
export files can be retrieved by your FTP client with the 'GET'
and 'MGET' commands.
If you'd prefer to use a local FTP server instead of RLG's as the destination
of these exported records, that's an option. Just send me your FTP server's
address and the local FTP account and password for write access and I'll
change the set up here.
Another PUT setup option is to create a single FTP file for each record that
is exported, rather than appending all exported records for each day to the
same file. Let me know if you'd prefer the single-record per file option.
We can also change the file name prefix (currently "PUT"), if you'd like.
Thank you for testing the new RLIN PUT command,
Date: Fri, 29 Jul 1994 06:06:37 -1000
From: "Bruce Washburn"
To: chantiny@uhunix.uhcc.hawaii.edu
Subject: Re: RLIN and Eureka access from Wyse 50 terminals
Martha,
I've worked on the terminal-type setting issue with the eureka-info
access method of connecting to Eureka. Unfortunately, it won't
be possible to include non-VT100 support in the eureka-info program
in the near term, since this uses network hardware on this end
of the connection that can only support VT100's.
Test accounts are in the process of being set up here. What would
be nice for testing would be if your test PAC menu could be modified
to include a scripted menu selection that would telnet to the
address "rlin.stanford.edu", and end the script at that point without
logging on. Then you and other staff evaluating RLIN and Eureka
could interactively type in whatever RLIN accounts are provided,
and also interactively type in the "call eureka" command with
appropriate terminal type settings, if needed. Is that a
possibility?
Date: Mon, 1 Aug 1994 17:45:10 -1000
From: Martha Chantiny
To: "Bruce Washburn"
Subject: Re: PUT command test access
In-Reply-To: Your message of Fri, 29 Jul 1994 09:06:30 -1000
Aloha Bruce -
I think I've got most of the basic steps figured out for the PUT and FTP
transfer - but the filename "ulogYYMMDD" is too long for our system - we
are limited to 8 characters or less. We can rename before the "get" but it
would be easier if the records were saved in a file named something
like UHYYMMDD - so we can easily tranfer it over.
Also, can you find out if the files are put in the "ulog" file in LC MARC
or OCLC MARC format? And what marks the end of one record and start of the
next?
We could probably figure it out by running various programs on the data
and looking to see what happens ... but it might be faster if someone
could just tell us.
Date: Mon, 1 Aug 1994 19:14:38 -1000
From: Martha Chantiny
Subject: [Re: Trial account]
To: bl.jag@rlg.stanford.edu, bl.btw@stanford.edu, bl.krt@stanford.edu, bl.ks
s@stanford.edu
Aloha Bruce, Karen, Judie -
The test account listed below appears to be set up for VT100 type
terminals ONLY. The screen is a mess as soon as one connects using
anything but a VT100 (which in our case is Wyse50's set up as Wyse50's).
I cannot find any way to enter a terminal type setting once connected.
Is there some way you can set up this account to prompt for terminal type?
Or preferably to be compatible with Wyse50's (the default terminal type
for the other test account Y6.E81 seems to work OK with our Wyse60's - at
least for RLIN(cat))
Date: Tue, 2 Aug 1994 06:58:13 -1000
From: "Bruce Washburn"
To: chantiny@uhunix.uhcc.hawaii.edu
Subject: Re: PUT command test access
Martha,
The records that are output with the PUT command are in RLIN
MARC format. If you have experience dealing with LC MARC and
OCLC MARC, the RLIN MARC format should look pretty familiar.
The RLIN MARC format is basically USMARC with some extensions.
The RLG document "The RLIN MARC Record: Description and
Interpretation", provides much more detail on this format and
would be useful to you. I don't know if it is being sent
along with other technical services startup materials. If not,
it can be requested from RLG's Distribution Services Center
(bl.dsc@rlg.stanford.edu).
The file names for the user log and the PUT files themselves
will be more than 8 characters long, at least for the near
term. The log file name probably could be reduced to 8
characters, but the PUT file names present more of a problem
since they need to include more information. If you are using
a DOS-based FTP client, you should still be able to Get the
files from our FTP server. All of the DOS and Windows-based
FTP clients I've used will automatically reformat the source
file name into something DOS can accept during transfer, or
give you the opportunity during transfer to specify your own
target file name.
The PUT records are written, one after another, to the target
FTP file. There is no "end of record" marker. The log file
indicates the length of the output record in bytes, so a process
can be developed on your system to extract records by reading
the appropriate length of each record, in bytes, in sequence.
Sites like Princeton, Harvard, and Stanford have written their
own utilities for taking these output files and converting them
into blocked input files for their IBM mainframe applications
(as they deal with straight FTP files of MARC data from a variety
of sources). If have some source code from Stanford that they've
shared with us to help other sites interested in developing
a similar utility, which I can send to you if you'd like.
Remember that a "single-record-per-file" output option is available
for PUT, if managing a multiple-record file presents problems
for you. In the single-record method, a separate FTP file would
be created for each record that is exported. Your FTP PUT profile
is currently set up for the "Multiple-record" method, but I can
easily change that to Single if you'd prefer.
Date: Tue, 2 Aug 1994 19:34:44 -1000
From: Martha Chantiny
To: "Bruce Washburn"
Subject: Re: PUT command test access
Bruce, et al.
We have successfully brought across 2 files, and by next week will be able
to tell if our MARC record processing programs can parse the records in
the files. Thanks!
Date: Tue, 2 Aug 1994 19:38:01 -1000
From: Martha Chantiny
To: "Bruce Washburn"
Subject: Re: PUT command test access
Aloha Bruce -
Thanks for the further info. Turns out the problem wasn't the long name,
but that I needed to specify that the file was in binary. Got it all
figured out today with Steven and Kevin's help and it looks very promising.
By the way - you know the MARC record already has the record length in the
first 7 bytes or so - so there shouldn't be any reason why we need the log
file. Is this something other sites use?
We are going to work on the assumption that the "put ... exp" file will
have all the record length info needed by our processing programs (that's
the way it has worked with other ftp'd MARC record sources).
Date: Tue, 2 Aug 1994 15:39:39 -1000
From: Martha Chantiny
To: adamson@uhunix.uhcc.Hawaii.Edu
Subject: RLG 'PUT" files
James -
The 2 files, named "put08194" and "put08294" are on your vol/subvol.
Please see if any of the know processing/loader programs can deal with
them - sometime in the next week or so (the vendor report is STILL #1).
They look like standard MARC records, all strung one after the other.
I will move each day's file to your subvol via FTP (remembering to type
"bin" first) so you should have a few different files to work with by next
week. I won't move the log files, since the record size is already
included in each MARC record.
Date: Thu, 4 Aug 1994 10:56:16 -1000
From: Bronwen Solyom
To: Martha Chantiny
Subject: Re: RLIN test
Martha,
Thanks for the info. I plan to start our test with the OCLC "no hits" as
these will likely be "few hits" or "no hits" in RLIN as well and since
these items are waiting for rapid cataloging anyway, it will be OK to pile
them up. If you can get them into the work (or live files) we should have
no trouble finding them because there would be no other records for those
titles existing in the file. We can monitor roughly how many titles we
have put at the end of the week and will keep you informed. If they all
disappear, we will have printouts so we can could download them again from
rlin if we needed to!
I'm glad of the chance to do this test. I'm finding the comparison with OCLC most interesting.
Date: Wed, 10 Aug 1994 18:38:50 -1000
From: James Adamson
To: lib_ids@uhunix.uhcc.Hawaii.Edu
Subject: *** RLG TESTING ***
If your not Technical Services then you may delete now.
-------------------------------------------------------
Beginning tomorrow afternoon August 11, 1994 from 2pm till 3pm
there will be no downloading.
This is due to the need for Systems to test the RLIN "PUT" command
and to make sure we can produce a readable MARC record in the
workfile.
Mahalo for your assistance.
Date: Fri, 12 Aug 1994 19:31:29 -1000
From: Martha Chantiny
To: adamson@uhunix.uhcc.Hawaii.Edu
Subject: RLIN files using READFLOP
James -
I tried to do your procedure on Friday night on the newest 3 RLIN "put"
files, but the READFLOP went into inspect every time. (I tried it 3 times
since you didn't tell me if there was a way to do more than one "put" file
at a time as the "input file").
The latest 3 files are bigger than the previous ones but I don't know if
that is the problem. Anyway, they are on $data7.ja as put81094,81194 and
81294 if you want to try it on Monday to see what the problem is.
Date: Tue, 16 Aug 1994 08:31:33 -1000
From: Gary Macmillan
To: bl.krt@RLG.STANFORD.EDU
Subject: DA TEST
Seem to have hit a major snag in our test of the PUT command. On Tuesday of
last week our Systems staff finally had time to patch together the software
to FTP the files and load them into our system, worked fine for thru Thursday
on small groups of records. Friday the system bombed on a large file of
records, and again last night we got 13 records loaded from a large file and
again it bombed. We have lost a significant amount of work and records. The
problem, which was communicated to Karen by Martha Chantiny of our Systems
staff last night (by e-mail), seems to us to be at your end. What do you
know of it? I have told staff here to cease the use of RLIN/PUT command
until further notice, it really is a waste of time since it is a gamble as to
just how many records we will get.
I would appreciate anything you can do to move a solution along.
Date: Tue, 16 Aug 1994 07:46:46 -1000
From: "Bruce Washburn"
To: chantiny@uhunix.uhcc.hawaii.edu
Subject: "Garbage" in PUT files
Martha,
I've looked at some of your output files, and it does appear as
though somewhere in the file an output record was over-written,
causing the sequencing using the MARC leader to be thrown off.
This is a serious problem, and I'm surprised no other site has
reported it. It did not occur during our extensive internal
testing of the PUT command last June, however there have been
a couple of RLIN system updates since the PUT command became
available and it is possible that a recent system update has
precipitated this problem.
I have a few questions that I need help with, as I pursue a
solution to this here.
1) Did you use the PUT command prior to 8/10 without
seeing this problem (your note suggested that you had
processed files successfully prior to the 10th, just
checking to make sure)
2) Is the PUT command being used by more than one HAUG
account simultaneously?
3) Can you process the output file for 8/15, and let me
know if and where it fails (the log file for this
has an inconsistency after record 53 in the log)
Date: Mon, 15 Aug 1994 15:28:15 -1000
From: Martha Chantiny
To: bl.kss@rlg.stanford.edu
Subject: "Garbage" in PUT files
Aloha Karen -
Things were going swimmingly with the FTP transfer and processing of "put"
RLIN records until we reached the file of August 10th. The
PUT.081094.EXP, 081194 and 081294 files all appear to have inappropriately
calculated records and/or "end of record" marks in strange locations
(strange from the point of view of our processing program at least).
We are sure it is not the act of using the "get" command or the FTP
transfer itself that is causing the problem. The problem seems to be some
kind of "buffer not clearing" situation. The examples, and order of the
files, that I list down below are based on the files located on your
rlgftp.stanford.edu in our directory (HAUG/ALOHA894).
Our processing program (called READFLOP, because these records most
closely resemble records which we receive from vendors on floppies)
uses the MARC Record leader to calculate the size of each MARC record
(since that is what the first 5 bytes of the leader is for). The
processing program has encountered "garbage" and choked on the
following records in the 3 PUT files referenced above.
081094 - the first 7 records are OK but the 8th one is preceded by a
couple of ranges of "random garbage" interspersed with end-of-file marks.
The last good record in the 8/10/94 put file is entitled "Buddhistische,
hinduistische und jainistische Skulpturen Bronzen und Ritualgegenst" and
has an ISBN of 3900950040.
081194 - the first 31 records are OK, but the 32nd one has garbage.
081294 - the first 13 are OK, but the 14th seems to start in the middle
of a record, rather than with the 000 leader tag and our processing
program reports errors such as "missing field terminator" and
"directory error". The last good record appears to be the one titled
"Ang lunes na mahirap bunuin", the LCCN might be 94-421 (no 010, this
info is from 008).
Since this is just a test, we are loading the records which make it thru in
one piece and ignoring the bad record and everything subsequent. If we go
into production mode, this problem would obviously have to be fixed. Has
anyone else had the same experience?
Do you have any suggestions for how this problem might be prevented for the
next 2 weeks? Other than having each individual record go into a separate
file (which would be more labor intensive from our system level processing
point of view).
Date: Tue, 16 Aug 1994 13:10:30 -1000
From: Martha Chantiny
To: "Bruce Washburn"
Subject: Re: "Garbage" in PUT files
* 1) Did you use the PUT command prior to 8/10 without
* seeing this problem (your note suggested that you had
* processed files successfully prior to the 10th, just
* checking to make sure)
YES, WE PROCESSED THE FIRST 6 "PUT" FILES WITHOUT ANY PROBLEMS (they were
all smaller). THE FILES FOR 8/2 THRU 8/9 WERE OK.
* 2) Is the PUT command being used by more than one HAUG
* account simultaneously?
YES, UNDOUBTEDLY. SEE OTHER MESSAGE.
* 3) Can you process the output file for 8/15, and let me
* know if and where it fails (the log file for this
* has an inconsistency after record 53 in the log)
ONLY THE FIRST 13 RECORDS OF THE 8/15 PUT FILE PROCESSED CORRECTLY. THE
GARBAGE OCCURRED ON THE 14th RECORD.
According to James, staff doing the "putting" identified the "garbage
record" based on what could be seen of it in the file as a record that had
been (successfully?) selected and processed on 8/12.
Could "debris" from former files somehow be infecting new files? I assumed
that each "put file" was created fresh by the action of the first "puttee"
and that each subsequent "put" was just concatenated to the top or bottom
of the file ...
Date: Wed, 17 Aug 1994 18:41:04 -1000
From: Martha Chantiny
To: adamson@uhunix.uhcc.Hawaii.Edu
Subject: RLIN put file(s)
James -
The 8/17 RLIN PUT file was "readflopped" without any inspect - 235
records. I added it to the RLIN file processed yesterday (about 8 records
I think) and so something like 243 records should print out on Thursday
when you do the printout for the department(s).
From: Martha Chantiny
Subject: [Re: PUT problem identified]
Thelma, Bron, Janette, Paul, Roberta, Kengo, Sharon:
Please divide the accounts and passwords (listed below) among the 3
departments depending on how many simultaneous staff are doing the RLG
testing (I suggest 4 accounts to Monographs, 2 to DBM, 1 to Serials and
leave the original one for Orig Cataloging - but you guys know best how
many SIMULTANEOUS testers there are in each department).
The problem with the bad "put" files is due to everyone using one account.
Rather than isolating each of the test lines into a separate service
(Test1, Test2, Test3, etc.) if you could simply agree to use different
accounts, and to make sure that no one else is using the same account
within your department - this should insure that future "put" files do not
suffer from "colliding records".
As long as everyone connected to RLG at any given time is using a DIFFERENT
account and password, there should be no more colliding records.
Date: Tue, 23 Aug 1994 18:25:58 -1000
From: Martha Chantiny
To: adamson@uhunix.uhcc.Hawaii.Edu
Subject: RLG records
James -
The combined RLG file from 8/22 ("ready") didn't work, so I individually
"readflopped" the 4 files in $data15.rlg from 8/22. I added the 4 files
>From 8/18 and renamed and transferred the 5 files from 8/23 to the
RLINUHL file before starting up the Broc processes again.
The files on the RLG machine didn't match in size for one of the 8/22
files on the Tandem, and I screwed up one of the files for the 8/18 load,
but as near as I can tell this is what was processed:
# of records
"PUT2" 8/23 - 14
"PUT3" 8/23 - 9
"PUT4" 8/18 - 111
8/22 - 87
8/23 - 31
"PUT6" 8/22 - 28
"PUT7" 8/18 - 2
8/23 - 1
"PUT8" 8/18 - 68
8/22 - 147
8/23 - 96
"PUT" 8/18 - 11
"???" 8/22 - 89
Which I think is somewhere around 694 records that should be in Live by
Wednesday morning.
Date: Wed, 24 Aug 1994 19:40:12 -1000
From: Martha Chantiny
To: paulb@uhunix.uhcc.Hawaii.Edu, janette@uhunix.uhcc.Hawaii.Edu, macmill@uh
unix.uhcc.Hawaii.Edu, kengo@uhunix.uhcc.Hawaii.Edu
Subject: Put records from 8/23
Paul, et al.:
The account BV.E19 with password 7xxxxx should have generated the PUT file
"PUT8" - which shows 96 records in it from 8/23 and which _was_ processed
on Wednesday morning. If you want to check the log entries (below) for PUT*
against what you THINK should have been downloaded on 8/23 and let Systems
know if you can identify specific items which you are SURE you "put" but
which do not show up on the log or in the processed files. (The log file
is generated by RLIN, and comes from THEIR computer not ours.)
There may be some problem with the PROCESS which handles RLIN records,
since I understand they are reporting that "regular downloads" aren't
showing up in the Live file today ...
Hmmmm.
Well, it IS only a test.
Here is the log of all files that RLG thinks were "put" on 8/23:
0823:090847: app PAUG86-S2236->PUT7.082394.EXP, 820 bytes
[...]
Date: Wed, 24 Aug 1994 21:45:05 -1000
From: Martha Chantiny
To: macmill@uhunix.uhcc.Hawaii.Edu, adamson@uhunix.uhcc.Hawaii.Edu
Subject: RLG stuff
Tonight I found only 2 "put" files for 8/24 - those were PUT8 (20 records)
and PUT7 (3 records). I processed those, and all are in the Live database
(I checked).
I then renamed that file "done824" and put the 8/23 RLIN file back in
place (on $data7.rlin). The large RLIN file is the one with over 600
records in it, the accumulation of 3 days of RLG "putting" plus regular
RLIN downloads of CJK records.
Before processing the 8/24 "put file" I determined that only about HALF of
the large RLIN file had been processed (using the printout provided by
Steven - which I have left on James's chair).
The processing apparently stopped at the 340th record in the file. This
is why everyone was reporting that the "put records" weren't in Live.
I have NO IDEA WHY processing aborted, but after stopping and starting the
BROC, the processing appears to have started up again and went for about
55 more records, then stopped again. (I've marked the records in the
printout.)
The sample record that Steven was given the title for is the second to the
LAST record in the file! So that explains why it is not in the Live
database yet.
On Thursday, it may be necessary to delete the first 390 records or so
from the RLINUHL file in order to get it to finish processing the rest
of the records.
If this is necessary, the all downloading will have to stop while the file
is fixed, the BROCs stopped and started up again.
To conclude, all the "put records" are ON our system, they just are not
all processed into the Live file, for reasons as yet unknown.
Date: Thu, 25 Aug 1994 09:49:33 -1000
From: Gary Macmillan
To: chantiny@uhunix.uhcc.Hawaii.Edu
Subject: [Martha Chantiny : RLG stuff]
Good detective work, there but not there, our problem or RLIN? Today is the
last day we will "put" until RLIN can correct the bug and have one, 8 digit
file. I will message them today about our cessation.
Date: Thu, 25 Aug 1994 14:55:54 -1000
From: Gary Macmillan
To: bl.btw@rlg.stanford.edu, chantiny@uhunix.uhcc.Hawaii.Edu, macmill@uhunix
.uhcc.Hawaii.Edu
Subject: ["Bruce Washburn" : PUT testing]
Thank you, Bruce, for your quick response. We will eagerly await the "fix on
put". Martha Chantiny will respond on Monday as to the file name format. I
believe that our Tandem has a limitation of 8 characters in file names, and
you use 12, the renaming of the files, etc proved to be more than the human
mind could conquer during this past week.
Date: Thu, 25 Aug 1994 13:40:58 -1000
From: "Bruce Washburn"
To: macmill@uhunix.uhcc.hawaii.edu
Subject: PUT testing
Kris Tague forwarded to me your note describing your experience with testing
of the RLIN PUT command, and you noted that, for this command to be of more
interest to you, you'd want the "multiple session" bug to be fixed and also
for PUT to create files with smaller file names. The PUT command bug that
surfaced in your testing earlier this month should be fixed soon after the
programmer who designed the FTP gateway is back in the office from his
vacation, after 9/6. There is an RLIN system software update currently
scheduled for 9/17, and our expectation is that this bug will be fixed,
tested, and installed in the production RLIN system on 9/17.
We may also be able to do something quickly to respond to the "smaller file
name" requirement, although I'd need to get a little more detail on what an
effective file name format is, from your perspective. Can you provide
some specifics on what the optimal form of an FTP export file name would
be, to fit into your processing?
Date: Sun, 28 Aug 1994 22:50:12 -1000
From: Martha Chantiny
To: Gary Macmillan , bl.btw@rlg.stanford.edu
Subject: PUT file name length
Bruce -
The Tandem will accept a "get" filename of
anything that is no more than 8 characters long - preferably without any
periods or other punctuation. For instance, I was renaming the RLG files
to "PUTmddyy" - this worked fine. If they all started with the month/day
then a MGET would work and bring across all files in one fell swoop.
Date: Mon, 29 Aug 1994 09:30:57 -1000
From: "Bruce Washburn"
To: chantiny@uhunix.uhcc.hawaii.edu
Subject: PUT file name length
Martha,
Thanks for supplying the additional detail on your preferences for
PUT file names, that's very helpful.
Date: Mon, 29 Aug 1994 19:20:21 -1000
From: James Adamson
To: kengo@uhunix.uhcc.Hawaii.Edu, macmill@uhunix.uhcc.Hawaii.Edu
Subject: *** RLG PUT STATUS ***
The RLG PUT is now complete and the MARC record report is on
Kengo's desk.
Out of 238 records that were completed. (1) was corrupted, which
happened to reside in a "block" so approx 8 records were lost. The
corruption was contained in a bad "directory" or "map"; (2) approx.
5 records had blank 050's and therefore these records were lost;
and (3) all records [FYI] had a tag 001 control number with non-
digits.
So... 225 MARC records were processed from the final TEST run.
Ma halo Nui Loa for your patience and kindness.
Date: Thu, 10 Nov 1994 11:44:23 -1000
From: "K. Smith-Yoshimura"
To: macmill@uhunix.uhcc.Hawaii.Edu
Subject: Shorter FTP PUT File Names Update
Dear Gary,
Our technical staff have verified that shortening the file names of records
"PUT" to your FTP server so Tandem doesn't "choke" does not represent a major
task, and they have begun working on it. However, we have a deadline to
implement Format Integration for variable fields by 12/31/94 that cannot be
changed, and we do not know for sure if we can implement the FTP file name
change request at the same time as Format Integration. If we can fit the
change request in so that it is also implemented by 12/31/94 (with the Format
Integration variable fields implementation), we will do so; otherwise we would
implement it in the next RLIN system upgrade expected by early February 1995.
Date: Mon, 21 Nov 1994 14:42:25 -1000
From: "Bruce Washburn"
To: chantiny@uhunix.uhcc.hawaii.edu
Subject: RLIN PUT Command: Enhancements to FTP output file name options
The RLIN system has been updated to support a much wider variety of PUT file
names. These changes are a direct result of the input RLG received during the
pilot phase of PUT testing prior to September 1994.
You can now obtain greater control over the output file names created by the
PUT command, specifying an output file name of a length and composition
determined to meet your system requirements. And you can also take advantage
of a new automated file name component to differentiate between bibliographic
and authority file output.
More details on the new file name options are given below. The PUT
documentation will be updated shortly to reflect these changes, and I'll alert
you when a new copy of the documentation is available. With this increased
flexibility on the part of the system there is increased complexity in the
process of sorting through all of the different file name options. So if you
have questions on how to create the kinds of output file names you want, please
let me know.
Date: Tue, 22 Nov 1994 07:54:44 -1000
From: "K. Smith-Yoshimura"
To: macmill@uhunix.uhcc.hawaii.edu
Subject: Final RLIN PUT Barrier Removed: Enhancements to FTP output file name o
ptions implemented
Dear Gary
holiday and saw your e-mail -- hope your holiday was enjoyable.)
I've asked Bruce Washburn to set up a one-day "free RLIN PUT" test for Martha.
I trust you or Martha will let us know if "RLIN passes the PUT test" so we can
complete the invoicing/paperwork needed.
Glad we were able to surprise you with an early completion of the PUT changes!
Date: Thu, 1 Dec 1994 19:28:11 -1000
From: Martha Chantiny
To: "Bruce Washburn"
Subject: Re: PUT testing
* FTP server (select one)
IF THIS MEANS WHERE TO PUT THE SELECTED RECORDS ...
* XX___ Use RLG's
*
* File Name Template
ASSUMING THAT IT _WILL_ BE POSSIBLE FOR EVERYONE TO USE ONE ACCOUNT AND
SEND MULTIPLE RECORDS SIMULTANEOUSLY TO ONE "PUT FILE" ...
* Enter a default template only or both a default template and template for
* specific accounts. See the section "Specifying a Filename Template"
* below for more information.
*
* Default template:
PUT
MEANING WE WANT 2 FILES PER DAY, LABELED WITH ONLY MONTH/DAY but NO YEAR;
one file with BIB records, one file with AUTH records. So if I got 2
files tomorrow - they would be "PUT1202A" and "PUT1202B". I don't
_think_ it is a problem if the extension ".EXP" is added - but we
don't need it at our end.
IS THAT OK??
*
* RLIN account _?????? Filename template _________________________
DON"T YOU HAVE TO SUPPLY ME WITH THE RLIN TEST ACCOUNT???
Date: Fri, 2 Dec 1994 07:44:03 -1000
From: "Bruce Washburn"
To: chantiny@uhunix.uhcc.hawaii.edu
Subject: Re: PUT testing
Martha,
Test access to the PUT command has been established again for you.
Use the RLIN account BV.E13 (password xxxxxxx). This account
will allow multiple simultaneous RLIN sessions to be established
with this same account and password.
Files you create with the PUT command will be stored on RLG's FTP
server. To get these output files, FTP to
rlgftp.stanford.edu
and logon with the user name
HAUG
and the password
AXXXXXXXXXXXX
Your output files will be created using the filename template
PUT
As I noted in my earlier email response, the system-supplied date
needs to include the year, not just the month and day.
So output files created today for records transferred from the
bibliographic files will be in the file
PUT120294B
and records transferred today with PUT from the authority files
will be in the FTP file
PUT120294A
I've tested this with multiple sessions in RLIN, transferring
records to the server, and have had no problems with record
transfers over-writing one another. This problem, which you
encountered during the beta test period for PUT in August, was
corrected in early September. My testing today was just to
make sure that this problem is still fixed :-)
Date: Mon, 5 Dec 1994 13:09:52 -1000
From: Martha Chantiny
To: "Bruce Washburn"
Subject: Re: PUT testing
* Re: PUT testing
* We can set this up as you've requested, except for the date format
* you wanted (, without the year). The year needs to be
* included in the system-supplied date. I will provide you with a
* test account set up with the file name template, with the modification
* of the date being represented in form. If that is a problem,
* please let me know.
Bruce - if the file names are "PUTmmddyyA.EXP" and "PUTmmddyyB.EXP" then they
will be TOO LONG. So unless you are saying that the file name will be
"Ammddyy" and "Bmmddyy" - we will have to try another format. Thanks,
Date: Thu, 8 Dec 1994 13:03:12 -1000
From: Martha Chantiny
Subject: RLG "Put" testing
A temporary menu item has been added to the POA menu - for testing of the "new
improved" RLG PUT command.
In theory, many people should be able to "put" both Cataloging and Authority
records simultaneously into just one file per record type. AND those files
are supposed to be named with only 8 characters (which our Tandem requires).
My understanding is that full scale testing will not start until next week -
due to all the events this week. However, if you take a look this week and
find that anything is really weird - please send a message to 'request'.
The menu option is set up to automatically connect to the new test account
that RLG has provided, and from what I can tell, it works with WYSE50
terminals OK. However, I don't have a clue about how to search, or "put" ...
Date: Fri, 9 Dec 1994 15:19:11 -1000
From: Kengo Yasumatsu
Subject: Testing RLG "PUT" COMMAND
******************************************************************
ONE DAY TESTING OF RLG'S IMPROVED "PUT" COMMAND ON DEC. 12, MONDAY
******************************************************************
I am asking POA cataloging staff who have access to the WYSE 50 terminal
to test RLG's improve "PUT" command to obtain RLG records on DEC. 2,
MONDAY. This is a one-day test to see whether RLG has corrected problems
we had encountered in using the "PUT" command to obtain RLG records during
the last trial period.
Some of you already have received and read Martha Chantiny's message
regarding a new set up to connect to RLG - I am appending Martha's e-mail
at the end of my message.
=========================================================================
If you choose "23 ***TEST RLG CATALOGING CONNECTION***" on the POA menu,
you are connected to RLG and can begin to search in the RLIN database.
You need neither authorization numbers nor passwords to get into
the RLIN database according to the new setup!
=========================================================================
Please try to get as many records as possible and as many MARC formats as
possible to test a simultaneous use of the "PUT" command. Martha and
Systems staff need to verify RLG records being obtained no longer collide
each other when a simultaneous use of the "PUT command takes place at
various WYSE 50 terminals in POA.
According to RLG, we should be able to obtain CJK records, including
vernacular data, on the WYSE 50 terminal with the "PUT" command. I would
like to ask Hisami, Soonyoung, Sophia, and Suzie to obtain authority
records from RLIN using the "PUT" command so that Systems will be able
to verify that authority records will get into a separate file.
PLEASE NOTIFY SYSTEMS IF YOU ENCOUNTER ANY PROBLEMS IN CONNECTING TO
RLG THROUGH THE SETUP.
Date: Mon, 12 Dec 1994 20:03:43 -1000
From: Martha Chantiny
Subject: RLIN PUT Test
I "got" the "put" RLG Test Bib file and processed it into the Tandem
RLIN file, and all 458 records appear to have completed indexing as of
approximately 7:30pm.
The processing aborted near the end (sometime after 6:30pm) on the
record "Flood frequencies for selected streams in Hawaii" - but I was
able to restart the indexing, and then recreate the piece (display record)
that got lost in the abort.
As far as I can tell all records made it thru in one piece, into one file.
To all who "put" records on Monday - Please check your materials
against the records in the LIVE database, and let me know if the test
appears to have been successful from the POA side.
Date: Mon, 12 Dec 1994 20:11:08 -1000
From: Martha Chantiny
To: adamson@uhunix.uhcc.Hawaii.Edu
Subject: RLIN PUT test
James -
FYI - I used ftp to 'get' the files from the RLG FTP server.
I then used READFLOP to convert them into a file called RLINEW (created
using 'set like $data7.rlin.RLINUHL' in FUP).
I then stopped the BROCs and renamed RLINEW to RLINUHL.
I then started the BROCs again and began checking records.
About 1 1/2 hours after replacing the old RLIN file with the new one,
$BROCR went into Inspect.
I figured out (from the printout - which I am going to dump on Kengo on
Tuesday morning - all 500 pages of it) which title the processing had
stopped on - then stopped and restarted $BROCR and it picked up again with
the next title and apparently continued on without any problem until the
end.
If you want to examine the ftp'd original - it is $data7.ja.B121294, the
readflopp'd version is now called $data7.rlin.rlgtest. The BID of the
record that bombed is 0,34,76,220. I couldn't see any reason why. It is
about the 400th record or so in the file.
(I stopped the BROCs and put the original Monday RLINUHL file back in place
- I don't know if I needed to or not but it is all back the way it was
before the test.)
Date: Tue, 13 Dec 1994 08:02:06 -1000
From: Kengo Yasumatsu
To: chantiny@uhunix.uhcc.Hawaii.Edu
Subject: RLG "PUT" COMMAND TEST
Martha - Thanks for the RLIN PUT test message. It appears that the test
was success so far. I received several reports that the RLG connection did
hang. I observed one hung screen after the PUT command was typed at the
command prompt on screen. After breaking off from the hung screen, the
cataloger was able to complete the PUT command transaction for the same
bibliographic record. I also received comments that responses to various
commands were very very slow at some points!
Date: Tue, 13 Dec 1994 12:27:18 -1000
From: Martha Chantiny
To: karenf@uhunix.uhcc.Hawaii.Edu
Subject: Ops Manual addition
Karen -
If there is not already a document on "Processing FTP'd RLIN records" -
please create one, and put the following information in as a "placeholder".
>From TACL telnet to
rlgftp.stanford.edu
- or -
36.26.0.15
Logon with user name (in caps) HAUG
Password (in caps) AXXXXXXXXXX
Type "BIN" at the ftp prompt to set the file type to Binary.
Check file listing - there may be 3 files: AMMDDYY
BMMDDYY
(where MMDDYY = monthdayyear)
ulogMMDDYY
We do not need the "ulog" file.
GET the Ammddyy and Bmmddyy files - send to $data7.rlin.
Exit from the FTP connection by typing "bye".
Using FUP create a new RLIN file (rlinew), using "create like RLINUHL" on
$data7.rlin.
Run $carl.load.READFLOP on the Bmmddyy file - the output file is the new
RLIN file.
Obey $carl.broc.stopbroc
Rename the old RLINUHL file to RLINUHLX.
Rename RLINEW to RLINUHL
Obey $carl.broc.strtbroc
Back to Martha's Homepage