[FoRK] [fonc] Software Crisis (was Re: Final STEP progress report abandoned?)
meltsner at alum.mit.edu
Mon Sep 9 11:02:26 PDT 2013
On Mon, Sep 9, 2013 at 12:28 PM, J. Andrew Rogers <andrew at jarbox.org> wrote:
> On Sep 9, 2013, at 6:07 AM, Gregory Alan Bolcer <greg at bolcer.org> wrote:
> > That's what "Lean" startups are all about, though I'll probably get shot
> for criticizing the single most popular software management technique on
> the planet. It's like Agile, it's a technique that's great when properly
> applied, but many wield it like a magic hammer without truly understanding
> when it's applicable and when it's not.
> I was going to write something similar. The "lean" method selects for
> minimum investment in quality software engineering. It is a reasonable
> choice if your Big Idea is a photo sharing app for lichen enthusiasts but
> you cannot build a high-value technology base in this way. The focus on
> lean startups goes a long way toward explaining why Silicon Valley is
> producing a dwindling number startups producing anything resembling
> "technology", either directly or indirectly. The business models have been
> completely stripped of technology development because it is viewed as
> unnecessarily expensive in this model.
> Just looking at software, there are many startup ideas that can't even get
> off the ground for less than a hundred thousand of lines of code.
> >> In other words, I don't believe
> >> there's a software crisis or any real shortage of programming talent (I
> >> know plenty of great programmers who regularly go without work, often
> >> because they're unimpressed with the offers they're seeing.)
> Where are all these "great programmers" when you want to hire one? And
> exactly how large an offer is required so that they will deign to work for
> you? Around here, the going rate for great software talent is already
> creeping north of $200k before incentives and even the merely above average
> are far in excess of $100k.
> To a first approximation, there is quite obviously a fixed supply of
> talent and a rapidly growing demand. Increases in the clearing price are
> just redistributing the shortage.
It's tricky -- I think employers often have very different pictures in
their minds when they think of "talented programmers."
True story, somewhat scrubbed of detail. Long-standing customer, a
government IT department, was considering dumping a CASE (upper-CASE --
fully integrated software engineering environment) in favor of Visual
Studio. My acquaintance, a consultant with great expertise in the CASE
product, asked them why, given that the tool allowed a single programmer to
produce and to maintain as much software as a dozen programmers with Visual
Studio. The customer responded, "But it's really hard to find that single
programmer -- we can hire Visual Studio programmers by the dozen." And
went on to talk about the difficulties of managing unique individuals, the
relative pay scale (and ease of replacement), etc.
>From the customer's point of view, the high productivity was less important
than the rarity of individuals with those skills. Sometimes, it's easier
to go for the popular/mediocre instead of the highly productive and
esoteric -- companies are usually more comfortable with programmers as
fungible commodities instead of rare, special snowflakes.
Given the sort of programming I've seen in the wild, I think the talent
pool is exceedingly shallow when it comes to great programmers, but that
most employers manage to get by with less-awesome individuals, in part by
dumbing down the problems, taking on long-term maintenance headaches, etc.
In fact, I suspect many IT departments have never dealt with a really good
system as a result -- mediocre and buggy is the norm.
More information about the FoRK