# C[omp]ute

Welcome to my blog, which was once a mailing list of the same name and is still generated by mail. Please reply via the "comment" links.

Always interested in offers/projects/new ideas. Eclectic experience in fields like: numerical computing; Python web; Java enterprise; functional languages; GPGPU; SQL databases; etc. Based in Santiago, Chile; telecommute worldwide. CV; email.

© 2006-2015 Andrew Cooke (site) / post authors (content).

## Reflections on First Consultancy Gig

From: "andrew cooke" <andrew@...>

Date: Mon, 8 Jun 2009 14:51:37 -0400 (CLT)

I finished my first job in the capacity of a "consultant" and have been
thinking about how it went (ie worrying over the problems).

Technically, things generally went well.  I perhaps didn't worry enough
(or early enough) about efficiency, but this is perhaps more a social
issue than a technical one: even if you know a system can be made quick
enough in later iterations, "it's slow" is an immediate reaction of anyone
who tries it.

The technical/social distinction continues to be misleading, because my
first social issue is also a technical one: that I didn't think about the
future of the system.  Who will maintain and run it?  And do they have any
interest in it succeeding?

Related to that, did I provide what the client wanted?  I am not sure I
did.  There was an existing application that looked very good, but had a
poor technical basis.  I developed a new system that provided the
technical infrastructure on which a replacement could be built (or to
which the existing system could be moved).  That's fine, technically, but
it would have been much better received, I think, if it had been
implemented as a progressive extension to the existing system.  An
adaptive approach like that would have had the following advantages:

- The final result would integrate the existing presentation layer, which
looked good.

- During development I would have seen "real life" loads from the upper
layer, which might have led to slightly different (efficiency related)
implementation choices.

- The process for supporting the existing system would have adapted
gradually to the new system.

So why didn't I do this?  One obvious reason is that it sounds like a lot
more work.  But it turned out that we had enough time to implement much
more than anyone originally hoped.  So, at least in retrospect, I could
have taken much longer - half the functionality, better received, would
have been a decent tradeoff.

Another reason is that there was both poor communication and a fear of
poor communication.  The approach sketched above requires good
communication, I think.  But perhaps pressing forwards would have improved
communication to the degree necessary?  It is hard to argue against the
alternative - that by striking out alone, I made communication even worse.

All above is particularly galling because, while I don't think I am the
world's greatest communicator, I do think I am aware of my limitations and
sensitive to the kinds of issues described above.  I was certainly aware
of problems half-way through, when we gained an extension for more work;
at that point I knew I was not involving the client, but had not grasped
the extent of the problem.

While my *internal* process might be described as agile - many small
cycles, code always close to deployable, etc - I failed to implement what,
to me at least, seems to be agile's greatest tool: client participation.
And the key to client participation, in this case, would have been
adaption of the existing system.

Andrew