[Archivesspace_Users_Group] Implementing PUI

Thu Nov 15 14:30:06 EST 2018

Hi Christie,

We just did this here at GW: https://searcharchives.library.gwu.edu.

It was a really good experience! We aimed for a minimal process, maximizing
as little staff time as possible.

For context and for those who aren't Christie (Christie is already very
familiar with our context!): We previously published EAD XML files by
putting them on our web server. They were linked to style sheets which also
sat alongside the XML files on the server. So, it was just files on a
server. Our catalog linked to those finding aids, as did our institutional
website. And we relied on web search engines to index and search the
finding aids. Two weeks ago, we replaced those finding aids with the PUI.

*Github public documentation*
As part of our PUI project, we documented our local configurations to the
ASpace app here: https://github.com/gwu-libraries/aspace-pui/. That repo
shows everything we customized in relation to the PUI.

*Here are some of our fondest memories and lessons:*
1. A big +1 to Mark's pre-launch checklist
<https://archivesspace.atlassian.net/wiki/spaces/ADC/pages/103526318/PUI+pre-launch+checklist>
recommendation
to "Review the use of the 'publish' checkbox in your staff site". This was
big work, for which we relied heavily on the API. We made lists of all our
resources, compared that to our list of published finding aids, and
manually reviewed everything that didn't match. This was great. We found
resource records that had never been published, but which we wanted
published. We did a light review of everything that would be newly visible
to the public, including a double check that there wasn't anything
embarrassingly "drafty" or marked as restricted on these finding aids. We
tried not to be perfectionists, and that worked really well. We probably
spent about three days reconciling the lists of ASpace resource records and
published finding aid files. We found that it was best managed with one
person coordinating and sending out lists to others to approve what should
be published and what shouldn't.
We also unpublished large swaths of material, although our choices here
were particular to our institution. We unpublished digital objects,
accession records, and all unused repositories. We did this using the API,
but database changes may have been easier.

2. Agents - People worried that our Agent records might cause private
information to get published, like donor contact info. They don't. We also
chose to unpublish our staff agent records (aka, our own names).

3. You will absolutely want to search for weird stuff like "test" and
"demo".  If you're like us, you'll find stuff. So much stuff. Sometimes
hard to find. Often with the "published" checkbox checked.
Our favorite:
[image: blahdeeblah.JPG]

4. We came to terms with the fact that our subject headings need a lot of
work. We weighed wether to fix them before going live or after, and decided
to do it after. Since we had flat XML finding aids before, our subject
headings didn't really matter that much; they weren't surfaced as facets.
We decided that we can live with less-than-ideal subject headings for now.
I expect many other institutions would make the choice to fix up their
subject headings before launching. If so, the API would come in very handy
here, too.

5. For now, we've done redirects from the old finding aids to the new ones
in ArchivesSpace. We did a lot of internal outreach to make sure that
people understood that the PUI still creates "finding aids"/"collection
guides", and that the old finding aid URLs have a corresponding new URL.

6. If your finding aids are being harvested into ArchiveGrid, check that
everything will transition smoothly. It may be an opportunity to change
your method of harvest. We left ours in place as-is, because it was pulling
from the MARC records that we send to WorldCat, and the ArchivesSpace PUI
doesn't impact that process at all.

7. One last warning: we had to do some clean up that involved moving
resource records between repositories. That went poorly, and broke stuff.
Things were very weird for a while, because our index broke and then we
couldn't determine if it was fixed or not for a while. Lyrasis was very
helpful during this period.

8. Because part of our backend system would now be public, we realized that
we needed to improve security on our accounts.

9. We archived the old finding aids in Archive-It, and are keeping a
(non-public) copy of the XML finding aids for one year.

10. We couldn't have done any of this without the leadership and vision of
our public services librarian (Leah Richardson!), who was in full support
of embracing change.

*Project progression:*
1. June: Two 2-week sprints to configure the PUI and do most of the
cleaning of our data, at the end of which we launched what we branded as
the "beta" PUI.

2. June - October: The "beta" PUI existed alongside our XML finding aids
for several months (the "summer", which ended up extending into October).
During that time, we got a feel for how much it would "cost" to maintain an
application that not only contains our staff system, but also now our
front-facing system (esp from a sys admin perspective). We weighed that
against the improvements that ASpace would bring us. This wasn't too
formal. Lots of talking and checking in. No metrics or charts. We decided
in the end to move forward. We also used this time to ask staff to use the
PUI instead of the XML finding aids, with two purposes in mind: (1) get
staff trained, and (2) locate and fix messy data (e.g., if they saw any
more "Mr. Blah Dee Blahs", everyone was supposed to fix it on the spot).

3. October: A 1-week sprint to finalize the implementation of the PUI,
replacing the XML finding aids. We finished in our Halloween costumes.
(Special thanks to one colleague in particular who stayed late, in
head-to-toe costume!)

Best,
Rachel and Leah

Rachel Trent
Digital Services Manager
Special Collections Research Center
Gelman Library,  George Washington University
2130 H Street NW
Washington DC 20052
202.994.4824
racheltrent at gwu.edu

On Wed, Nov 14, 2018 at 3:20 PM Christie Peterson <cpeterson at smith.edu>
wrote:

> Hi,
>
> We're looking at implementing the PUI much sooner than we had previously
> anticipated. I'm aware of this pre-launch checklist
> <https://archivesspace.atlassian.net/wiki/spaces/ADC/pages/103526318/PUI+pre-launch+checklist>,
> but I'm wondering if others who have tread this ground before us have other
> resources (project charters/plans, checklists, etc.)  that you could share
> with us.
>
> Many thanks in advance,
>
> Christie
>
> --
> Christie S. Peterson
> Head of Technical Services for Special Collections
> Smith College
> cpeterson at smith.edu
> she/her/hers
>
>
> _______________________________________________
> Archivesspace_Users_Group mailing list
> Archivesspace_Users_Group at lyralists.lyrasis.org
> http://lyralists.lyrasis.org/mailman/listinfo/archivesspace_users_group
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lyralists.lyrasis.org/pipermail/archivesspace_users_group/attachments/20181115/5e7d2d38/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: blahdeeblah.JPG
Type: image/jpeg
Size: 29119 bytes
Desc: not available
URL: <http://lyralists.lyrasis.org/pipermail/archivesspace_users_group/attachments/20181115/5e7d2d38/attachment.jpe>