<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
</head>
<body>
<p>Fellow maintainers,</p>
<p>I've been developing software for reducing and analyzing
astronomical data since I started my Masters thesis in 1973, now
pushing toward 50 years ago. For the first 10 years, I pretty much
had to write everything myself, including a few local subroutines
that other people had written, but usually having to rewrite them
to do what I wanted. As I adopted formats which were becoming
standard, I wrote my own libraries to implement them, occasionally
adding code others had written. Thirty years ago, I put out my
first open-source package, which fit into a big system which had
been in use for less than ten years at that point. I had been one
of the first users in the mid-80's and understood it pretty
deeply, but was not ready to maintain the entire system. Luckily,
a few others have been doing that, though that maintenance has
been fading as younger astronomers switch to Python. For me to
move to Python, an entire software ecosystem for dealing with the
type of data the current system handles transparently would have
to be created, so have not moved it yet. <br>
</p>
<p>When I saw a need for a second package in astronomy a few years
later, it turned out that a lot of the libraries I had already
written could be repurposed, and a few new subroutines written.
Over the past 25 years, I developed a self-contained package in C
which has gradually become robust to compiler changes as it has
gained users. The interfaces at both the top level and the lower
levels have remained *almost* unchanging. One way to find out
who's using the code has been to make a change deep in the program
and find out who responds. Someone always does. Luckily, it's
technical-enough software that I can publish a paper describing
what it does in pretty good detail, but I'm turning 69 in a couple
of weeks, and will someday retire (after I publish that big
as-yet-unwritten paper), and I'm worrying a bit about succession.
My generation has been so successful in encapsulating knowledge of
this sort in software that the next generations haven't needed to
understand the concepts well enough to write their own software. I
suspect this problem happens with most complicated software,
processes, and devices. We need to foster future tinkerers who
will be able to figure out what is going wrong when it inevitably
does.</p>
<p>-Jessica Mink<br>
Smithsonian Astrophysical Observatory<br>
Cambridge, Massachusetts<br>
<a class="moz-txt-link-freetext" href="http://tdc-www.harvard.edu/mink/">http://tdc-www.harvard.edu/mink/</a><br>
</p>
<div class="moz-cite-prefix">On 08/18/2020 10:13 AM, Kara Andersen
wrote:<br>
</div>
<blockquote type="cite"
cite="mid:4B2FC543-7B28-44BD-AD9E-5F28BC507A76@gmail.com">
<meta http-equiv="content-type" content="text/html; charset=UTF-8">
<div class="">Hello Maintainers -</div>
<div class=""><br class="">
</div>
<div class="">In case you haven’t seen it already, I thought I’d
share this XKCD comic about maintenance.</div>
<div class=""><br class="">
</div>
<div class="">Kara</div>
<div class=""><br class="">
</div>
<div class=""><img apple-inline="yes"
id="355BFD65-361A-4847-A4A5-EB554F429FE4"
src="cid:part1.9F397A87.39822986@cfa.harvard.edu" class=""></div>
<div class=""><br class="">
</div>
Permalink: <a href="https://xkcd.com/2347/" class=""
moz-do-not-send="true">https://xkcd.com/2347/</a>
<br>
<fieldset class="mimeAttachmentHeader"></fieldset>
<pre class="moz-quote-pre" wrap="">_______________________________________________
Themaintainers mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Themaintainers@lists.stevens.edu">Themaintainers@lists.stevens.edu</a>
<a class="moz-txt-link-freetext" href="https://lists.stevens.edu/mailman/listinfo/themaintainers">https://lists.stevens.edu/mailman/listinfo/themaintainers</a>
</pre>
</blockquote>
</body>
</html>