x-www-form-urlencoded prohibition

[Russ Allbery]

Hi Mark,

I've been putting off responding to this in the hope that I could find
some time to do a proper job of it, but I'm about to disappear on vacation
for a bit and didn't want to leave without at least saying how much I
appreciated your thoughtful response.

I think we're coming to some questions of long-term maintenance with very
different architectural assumptions, but yours come from vast experience
with astronomy specifically and mine come from other fields.  That matters
a great deal, and I wanted to say that I'm listening to the feedback that
you and others are listening very closely.  It may well be that some (or
all!) of my assumptions are incorrect.

I did want to be sure to reply to this part in particular because I wanted
to at least be clearer about what my intent is:

Mark Taylor <m.b.taylor at bristol.ac.uk> writes:

> This may look like an acceptable cost to the large well-funded projects,
> developing new software, that are the loudest voices in the P3T.  But
> legacy data providers, having fewer resources to think about these
> things, are necessarily less active in the discussion, and science users
> of the VO won't know about such changes until things stop working.  If
> the VO decides to go in this direction, I hope that the impact on those
> stakeholders will receive adequate consideration.

It's precisely because that well-funded status is temporary that I am
thinking so much about these design questions.  We may be a well-funded
project now, but today's well-funded project is tomorrow's legacy data
provider.  I know I'm reaching much different conclusions than you have
reached, but I don't think that's because we disagree about goals.  I am
trying to minimize long-term maintenance costs across all their varied
axes: people, software, services, security, portability, and so forth.

My hope is that well-funded projects can use that funding while we have it
to help astronomy as a whole, including less-active projects with fewer
resources, by investing our resources into improving both protocols and
the tools that implement them in a way that everyone can benefit from.  We
have a window where we have funding and can contribute back to the broader
community; inevitably, that window will close, and then it will be up to
the next project with a window of funding to take the next steps forward.
My hope is that we can find a model of both protocols and software that
allows each new project with a surge of funding to bring all of us with
them.

I am coming to this with a background outside astronomy, and the amount
that I do not know about the ins and outs here is vast.  It is entirely
possible that this is too idealistic of a goal or that I'm going about it
in entirely the wrong way.  But I wanted to make sure that no one thought
my goal was to leave anyone behind.  Quite the opposite.

-- 
Russ Allbery (eagle at eyrie.org)             <https://www.eyrie.org/~eagle/>