Archive - Originally posted on "The Horse's Mouth" - 2006-05-28 05:19:02 - Graham Ellis
Even before I started University way back in the 1970s, I was working as a tape librarian - a library of some 60,000 tapes of data recorded by boats and land crews undertaking seismic oil exploration - setting off explosives or using a strong vibrator to shake the earth and record echos from which the strata deep below could be mapped. A fascinating technology.
Yesterday afternoon, I was speaking with my modern day counterpart at Saudi Aramco, who is responsible for about 10 times the number of "field tapes" we had back in those days - and the tape's capacity has risen by several orders of magnitude too. But there's still the same issue of knowing where a tape is if needed, of keeping backup copies off site, and of managing the whereabouts of each tape plus the 10,000 or so extra reels (sorry - cartridges now) which are used for processing work-in-progress.
Some things remain the same, it seems - such as the need to be utterly sure that the data from the field does NOT get overwritten - it would cost a fortune to send a crew back to a surveyed area. And the need to ensure that the right process tapes are available at each stage along the way. The mechanisms may have changed, tape use may now be recorded automatically to a database of some sort, but there's still the issue of loosing tapes or the records of a tape along the way which means "dead" units in the system; too much of that leads to a squeeze on active / available resources that we were very familiar with way back when.
These days, there are some commercial TMS (tape management systems) available, and there's OpenTMS too. As opposed to the old system I once wrote, this latter is written, I understand, largely in Perl and hence my delegate's attendance on the course. Wow - what would I have given to have had Perl on those old Xerox SDS9300 computers. What took me weeks to develop in Fortran would have been a couple of days of work ... and all the extra admin and tracking facilities described to me could have been practically implemented too.