<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
</head>
<body>
<div class="moz-cite-prefix">
<p>Le 16/12/2020 à 15:52, Greg Troxel a écrit :</p>
</div>
<blockquote type="cite" cite="mid:rmituslref1.fsf@s1.lexort.com">
<p>It's not clear to me where on the path from concept to full
implemention the ISO registry is. At a quick glance it seems to
contain only a small fraction of what is in the EPSG database.</p>
</blockquote>
<p align="justify">My current understanding is that software such as
PROJ, Apache SIS, etc. would not be direct consumers of ISO
registry. Instead, ISO registry would feed EPSG (or other)
registry, which in turn feed PROJ (or other) software. My guess is
that the rational for this indirection level is as below:</p>
<p align="justify">EPSG collects data from various national
agencies, but this not always done by or under the supervision of
agency staff. Errors may happen, and the EPG database has many
deprecated entries (7% of all entries last time I checked – not
all because of errors) replaced by new entries when errors were
found. By contrast, all entries in the ISO registry are either
provided directly by national agency staffs, or reviewed by them
before publication. From EPSG perspective, I think that the ISO
registry fills a hole: it provides a central authoritative source
where EPSG can refer instead than hunting themselves for data all
around the world.</p>
<p align="justify">Last time I checked, I found a web interface and
API for querying specific CRS, but nothing for downloading the
whole database (contrarily to EPSG). It contributes to my
impression that software interested in the whole database are not
the primary target of ISO registry.</p>
<p align="justify"> Martin</p>
<p align="justify"><br>
</p>
</body>
</html>