Errands

Dear colleagues,

I maintain **Record Atlas** (https://recordatlas.org/), a free, non-commercial,
open-source index of where genealogical records are held — which archive holds which
ground, for which years, and what it costs a researcher to look. It holds no personal
data and its own datasets are published openly.

The atlas currently draws a blank over most of the Czech Republic: of roughly 4,987
Czech places it maps, almost none can yet say which archive holds registers for them.
That is a gap in my index, not in your holdings.

The Digitální badatelna's own help page explains that archival records are linked to
authority records **shared by all Czech archives through CAM, the Central Archival
Module**, and that PLACE is one of their classes. That is precisely the thing I need:
not the records, not the scans, but the list of places and which archive's holdings
they attach to.

I would be grateful to know:

1. Whether CAM's PLACE authority records — the place name, any former or German name,
   and the archive or fonds they are linked to — can be supplied as a file, or are
   published anywhere I have not found.
2. If not as a file, whether there is an OAI-PMH, SRU or REST endpoint I may read
   politely, and at what rate you would like me to read it.
3. Whom I should credit, and under what terms.

I want to be explicit about what I have not done. The Opava portal
(digi.archives.cz) is behind a bot check, and several Czech portals return 403 to an
identified crawler. **I have not worked around any of them**, and I do not intend to.
That is why I am writing rather than collecting.

If a file is easier than an endpoint, a CSV of place name, former name, district and
holding archive would be more than enough, and I would publish the result with your
attribution and a link back to the portal.

With thanks for your time,

David Defranceski
Record Atlas — https://recordatlas.org/

Gentili colleghi,

I maintain **Record Atlas** (https://recordatlas.org/), a free, open-source index of
where genealogical records are held. It is not a commercial service and publishes its
own datasets openly.

I harvest the Antenati catalogue through the open data the Ministry publishes under
CC BY 4.0 — the OAI-PMH set `pico_m-ANTENATI` at

    http://www.san.beniculturali.it/oaicat/OAIHandler

**That endpoint has stopped answering, and it may not be only the endpoint.** As of
12 October 2026, from this network:

| host | resolves to | answers |
|---|---|---|
| `san.beniculturali.it` | 2.42.228.36 | no — times out |
| `www.beniculturali.it` | 2.42.229.1 | no — times out |
| `cultura.gov.it` | 2.42.229.1 | no — times out |
| `san.cultura.gov.it` | no DNS record | — |
| `antenati.cultura.gov.it` | an AWS load balancer in eu-south-1 | yes |

Every host on your own network is silent to us, including the Ministry's main site,
while Antenati's AWS-hosted front end answers normally. That pattern reads either as
an outage on that network or as that network declining requests from this range — and
I cannot tell which from the outside, which is partly why I am writing.

I also checked whether the endpoint had simply moved with the migration from
beniculturali.it to cultura.gov.it: `san.cultura.gov.it` has no DNS record. And the
OAI-PMH registry at openarchives.org — discontinued in July 2025, but its October 2025
dump of 7,103 repositories is still published — has never listed SAN, so there is no
alternate base URL recorded anywhere I can find. Both of the Antenati datasets on
dati.gov.it point at the one address above, so there is no file to download instead.

I had harvested 6,000 of the 464,770 records before it became unavailable, at one
request every two seconds.

Could you tell me:

1. Whether the endpoint is temporarily down or has moved;
2. If it has moved, the new address;
3. Whether a bulk file (XML, CSV or a dump) exists that I could take instead, which
   would be far lighter on your servers than 9,000 OAI requests.

I should add that antenati.cultura.gov.it returns 403 to any identified crawler while
its robots.txt permits everything. **I have not worked around that**, which is why I
use the open data endpoint and why I am writing now that it is unavailable.

Italy is the largest gap in this index — some 8,547 places it maps cannot yet say
which archive holds their registers — so I would be glad of any route you can suggest.

Con i migliori saluti,

David Defranceski
Record Atlas — https://recordatlas.org/

Dear colleagues,

I maintain **Record Atlas** (https://recordatlas.org/), a free, non-commercial,
open-source index of where genealogical records are held — which archive holds which
ground, for which years, and at what cost to the reader. It holds no personal data.

I am writing rather than collecting, and I want to be plain about why.

**sok.riksarkivet.se/robots.txt names my crawler.** It carries
`User-agent: ClaudeBot / Disallow: /` and `User-agent: Claude-SearchBot / Disallow: /`.
I read that as the archive's clear answer and I have respected it: I have not read
your search portal, and I have not gone to riksarkivet.se or data.riksarkivet.se to
get the same material by a side door. An institution that has said no by name has
said no.

But a robots.txt governs crawlers, not correspondence — so I am asking a person
instead.

What I would like is modest and is not the records themselves:

1. A list of the **parishes (församlingar)** whose church books Riksarkivet holds,
   with the county (län) and, if it is easy, the year span held.
2. Whether such a list is already published as open data, in which case a link is all
   I need.
3. Whom to credit and under what terms.

Sweden is currently 2,363 places on this map with nothing said about where their
records are. A parish list would fix that at a stroke, and I would publish it with
your attribution and a link back.

If the answer is that you would rather I did not index your holdings at all, that is a
perfectly good answer and I will record it as such and stop asking.

Med vänliga hälsningar,

David Defranceski
Record Atlas — https://recordatlas.org/

Dear colleagues,

I maintain **Record Atlas** (https://recordatlas.org/), a free, non-commercial,
open-source index of where genealogical records are held — which archive holds which ground,
for which years, and what it costs to look. It carries no advertising beyond a single footer
unit, holds no personal data, and its own datasets are published CC0.

Part of it is a surname dataset: every spelling a surname takes and the countries it is
attested in, built so that a researcher holding a name and nothing else knows which country
to start in. It currently draws on four national registers — the Polish PESEL list, the US
Census 2010 surname file, Spain's INE census frequencies and Norway's SSB table — for a total
of about 480,000 surnames.

**The Netherlands is a gap I would rather ask about than work around.** CBS publishes no
surname frequencies, and the Nederlandse Familienamenbank is plainly the authoritative Dutch
source. I have not scraped it and do not intend to.

I would be grateful to know:

1. Whether any extract of the familienamenbank — even name and frequency alone, without the
   etymological and distribution material — may be reused in an openly licensed dataset, and
   under what attribution and licence conditions.
2. If not CC0, whether a CC BY extract would be acceptable; I would then hold it separately
   from the CC0 file so the attribution requirement travels correctly rather than being lost
   in a merge.
3. Whether there is an existing route for this I have missed.

I am happy to credit the Meertens Instituut and the CBG prominently, to link to the
familienamenbank from every Dutch surname, and to exclude anything you would prefer not to see
redistributed.

With thanks for the work itself — it is the best of its kind in Europe.

Yours sincerely,
David Defranceski