26 Oct
2010
26 Oct
'10
1:27 p.m.
I'm getting into this a bit late (I was travelling last week), but I have a few
comments, based on 18 years of working at the Field Museum of Natural History in Chicago.
Ah no, you misnderstood me...
Presumanbly there is a database of the artefacts in the museum, indexed
by the indentification numbers. That database includes more details about
the particular object, ?things like (I would hope), options, serial
number (s), version, etc. All we've been discussing. I see no reason why
the date of acquisition (full date, not just the year), source (maybe
'anonymous donor' :-)), and the like should not be stored there as well.
It's far better to store too much information than too little.
But what I am wondering is why the year of acquisition should be a field
in the indentifier.
The Anthropology collections at FMNH were numbered sequentially, which was a bit of a PIA.
Having looked in the catalog ledger books, I might know that artifact 186276 is a Nazca
pot collected by Alfred Kroeber in 1927, but is 186277 from the same collecting field trip
or is it something from Tibet? Having a number like 1927.23.49 for that pot would make it
easier to find associated artifacts.
For my field work, I used RegionSite-Feature.artifact, such as M10-5.26. In the end,
though, any cataloging system is arbitrary.
<snip>
Message: 4
Date: Sun, 24 Oct 2010 19:22:50 +0100 (BST)
From: ard at p850ug1.demon.co.uk (Tony Duell)
Subject: Re: Cataloguing in a museum setting [was Re: nonsense...]
There could be an arbitrary limit because that is what some database programmer decided
was sufficient when the collections program was written (cf. Y2K).
On Sat, 23 Oct 2010, Tony Duell wrote:
That's _exactly_ my point. Having atbitrary limits may cause problems
later on.
Sure, but that's then 4 levels. I ahve no
problem extending the
heirarchical system to as many levels as are necessary, my query is why
it's noramlly limitied to 3. Why not just have as many levels as are needed.
A properly designed system should be extensible to as many levels as are
needed. And why recorsd the year of acquisition? What
importance is that? Why not
just a number for each artefact starting at 1?
It is unlikely that you nor I would care much about the year of
acquisition. ?But the bean-counters care.
-tony
I would also second the comment (Fred's?) that cataloging needs to be done by someone
with knowledge of the material being cataloged.
As to cataloging individual boards within a computer, we did not have a comparable
situation -- the FMNH's artifacts are usually single items. Where there were, for
example, multiple panels in a carved wall, the panels could be numbered 98123.1, 98123.2,
etc. The question that has been raised in regards to computers is, where do you stop
dividing the artifact??Again, the decision is somewhat arbitrary. The boards might not be
numbered, but presumably there is a description of the computer on record?that lists its
major components.
Bob
28 Oct
28 Oct
11:23 a.m.
feldman.r at comcast.net wrote:
The Anthropology collections at FMNH were numbered
sequentially, which was
a bit of a PIA. Having looked in the catalog ledger books, I might know
that artifact 186276 is a Nazca pot collected by Alfred Kroeber in 1927,
but is 186277 from the same collecting field trip or is it something from
Tibet? Having a number like 1927.23.49 for that pot would make it easier to
find associated artifacts.
Why does that need to be "encoded" in the artifact number? Why isn't there a
"field trip ID", and to find anything from a given field trip you just search
for everything with the same ID?
I think a lot of these systems fall down because they try to put too much
meaningful information into one single ID, and that doesn't scale well as the
system grows - they'd be better off having an ID linking a single item to a
database record, and other fields in that database record may be used to tie
that individual item to related ones.
In a computing context, I hit the same issue of trying to decide what level to
catalogue at - e.g. should I catalogue every board in a system with a separate
database entry because capturing that data might be important? But taken to
extremes, if I did that does it mean I should be recording every IC on every
board because that S100 system in the corner over there happens to be the only
one with an example of serial board "foo" which uses IC "bar"?
There's a bit
of a "where does it end?" aspect to it all :-)
I kept finding myself leaning toward just doing things at the "system box"
level along with taking of extensive photographs (and listing of internal
cards and "important stuff"*), and if a researcher wanted more detail than
that then they'd just have some digging to do (which is fun anyway).
* as was said, whoever does the cataloguing needs knowledge of the item being
catalogued, and it shouldn't be down to just one person (the same thing is
true when it comes to deciding what items should be accepted into the
collection - the field's just too diverse for it to be down to one person, and
an item which seems dull to one member might be recognised as significant to
another).
And don't even get me started on how to try and organise stuff in storage... ;-)
cheers
Jules
12:28 p.m.
New subject: OT Re: Cataloguing in a museum setting
On 10/28/10 8:23 AM, Jules Richardson wrote:
should I catalogue every board in a system with a
separate database entry
CHM uses lot numbers (Xyyyy.nnnn) for artifacts that come together as a single donation,
and 9 digit accession numbers for each part that can have a separate storage location.
Sub-assemblies of an artifact with a single location are recorded in the descriptive
comment field.
--
The CHM inventory started as a Filemaker data base.
We now are using Mimsy XG
http://www.selagodesign.com/portfolio/mimsyxg.php
which is designed for museum collection management.
If someone were starting today, they would probably just use a relational data base with a
single
unique identifier and have related fields for the lot and assembly heirarchy.
You need at least the donor/lot grouping, and physical location of the artifact.
Collections management
needs several other groupings as well (on loan, etc.)
12:57 p.m.
New subject: OT Re: Cataloguing in a museum setting
Al Kossow wrote:
If someone were starting today, they would probably
just use a
relational data base with a single
unique identifier and have related fields for the lot and assembly
heirarchy.
Yes, that seemed logical to me - it just didn't seem to be the way that most
museums operate. Maybe it's down to historical hang-up (after all, I can see
that there are benefits to attaching semantics to ID numbers if you don't have
a nice computer-driven search system on-tap), or maybe there's some sort of
fundamental assumption made that *any* museum is like a library (just not for
books) and so has to have some kind of dewey-a-like cataloguing system.
You need at least the donor/lot grouping, and physical
location of the
artifact. Collections management
needs several other groupings as well (on loan, etc.)
Yes, I ended up needing quite a number of fields per asset record - some of
those being pointers to other records in other tables, each with several
fields of their own. I don't think that the implementation is by any means
complicated, but the amount of information that needs to be captured gets
quite large, and there's a lot of effort required in deciding exactly what to
capture.
It's surprising how much of the data ends up being quite freeform in nature
too, with lots of care needed when entering it into the system.
cheers
Jules
5023
days inactive
5025
days old
3 comments
3 participants
participants (3)
-
aek＠bitsavers.org -
feldman.r＠comcast.net -
jules.richardson99＠gmail.com