Assorted replies and acknowledgements
-------------------------------------
Tomasino wrote about tabs vs spaces[1], and tfurrows weighed in on the
subject too[2]. I am a bit of a heretic about this. I am primarily a
Python programmer (although, at least in principle, I am supposed to
be shifting effort to Lua and/or Go to better prepare myself for
programming in confined spaces. Contrary to what Tomasino said, the
Python language spec doesn't require spaces - the interpreter will
happily accept tabs, and will even let you mix tabs and spaces within
a file, so long as you don't mix them on the same line. However, the
official style guide declares spaces to be the preferred good style,
and the community is pretty fanatical about enforcing this. You will
be a pariah in the Python FOSS community if you insist on using tabs,
and so 99% of code does not.
Personally, I think this is totally backward and if I had the power, I
would convert the entire Python community to using tabs. The main
reason is that it is the default of every sane editor on every
platform from any point in computing history that pressing the "tab"
character (which is, of course, how every sane programmer indents
their code) generates - what else? - a tab? Getting the "correct"
Pythonic behaviour requires configuring your editor to act in a
non-standard way, ideally only for .py files. Now, this is easily
done in any text editor worth using, but it is problematic in that it
is very easy to not set this up *everywhere*. You can (and I have),
write a Python script on your main development machine, where your
vimrc is all happily pimped out for producing lovely .py files,
deploy that script on a server somewhere and, later, while ssh'ed into
that machine, quickly and unthinkingly make a small tweak/fix in
default vim, thus introducing tabs amongst the spaces. You can't see
that this has happened, until your code starts acting screwy because
it isn't indented as many characters as you think it is. This can be
a nightmare of a bug. And obviously in a Real Proper Development
Environment with Serious Business apps you don't do this, debugging
directly on the production server with an uncustomised editor. But in
a hobby project environment, everyone does this from time to time. It
shouldn't really be that big a deal. But with Python, it is, and I
hate it.
Any programming convention which results in source files which can be
subtly and invisibly corrupted by editing them with a default install
of either of the two most popular editors in *nixdom is ipso facto a
*bad convention* and should not be adopted despite any other arguments
in its favour. Sorry, Python community, but this is just obviously
true.
tfurrows wondered if one approach was more portable than the other.
I'd be happy to learn otherwise, but I am pretty sure there is no
difference - they are both standard ASCII characters.
yargo[3] and tfurrows[4] wrote about decentralisation of the public
access unix scene, and tfurrows mentioned "SDF gets a lot of
participation, but it's hard to get participation on some of the other
systems out there". This is very true. SDF is, as far as I know, by
far the most active and best-known of all the public access systems,
and I am pretty sure that the imbalance is totally out of proportion
to the quality of the systems. SDF is by no means terrible, but it
has its fair share of problems and I have a hard time believing that
every other system is far worse. Most likely, what is responsible
here is the fact that PAU systems unfortunately tend to share the
"network effect" that drives commecial social media. Things like
BBOARD and COM are only visible to other users of the same system (I
have written about this before[5]). If you join a system with a large
userbase, you can communiate with a lot of people about a lot of
things. If you join a smaller, newer system, you will have nobody to
talk to. This leads to a "rich get richer" effect, where people
preferentially join large, established communities even if arguably
superior smaller communities exist. This is a *very* powerful effect.
Even Google, with absurd amounts of money and brainpower, couldn't
beat this effect displace Facebook. I think it's unfortunate that our
systems have this effect in common, I think it hampers
decentralisation, but I don't know what to do about it.
Finally, tfurrows wrote[6] about disk encryption, in the context of
imagining "Big Brother" accessing his files. I think it's a genuine
tragedy that these kinds of thought experiments can no longer be
dismissed off the cuff. Once upon a time, if the government wanted to
spy on you, they had to send goons to stealthily break into your house
and plant bugs or cameras. Then men with sunglasses in large white
vans had to sit on the street out your house to receive the signals.
This kind of spying doesn't scale well. You have to buy bugs and
vans, and pay your goons salaries. Spying on twice as many people
costs roughly twice as much. Any intelligence agency with a finite
budget has a limited number of people they can spy on per year. If
the agency is rational, they will target "important" people (according
to some notion of who they "should" be spying on, whether you agree
with that notion or not). Unless you are involved in heavy crimes or
perhaps are an active political dissident, basic economics is enough
to very strongly suggest that you are not being spied upon.
All of this logic is out the window today. Most people voluntarily,
and at their own expense, surround themselves almost constantly with a
significant number of internet-connected devices running closed source
software and equipped with cameras and/or microphones. We are footing
the bill for our own bugs. Yes, the spooks have to find a way to
defeat whatever security these devices come with, and that might be
very hard and expensive. But once it's done for a common enough
device, it can be applied to millions of targets at very little extra
cost. It's actually economically viable to spy on a non-trivial
number of "complete nobodies", just in case. This *could* be done, so
it probably is.
Whereas once you had to be a little irrational to believe Big Brother
was watching little old *you*, nowadays you have to be pretty
technologically illiterate to think in the same way. We've ended up
in a world where the more you know about computer security, the
*easier* it is to convince yourself a vast surveillance apparatus is
bearing down on you. That's a terribly psychologically unhealthy
world to live in, IMHO.
[1]
gopher://gopher.black:70/1/phlog/20180729-tabs-vs-spaces
[2]
gopher://sdf.org:70/0/users/tfurrows/phlog/2018/ah6_spacesAndTabs.txt
[3]
gopher://circumlunar.space:70/0/~yargo/clog/zn-on-time-flu.txt
[4]
gopher://sdf.org:70/0/users/tfurrows/phlog/2018/ah6_redundantPubnix.txt
[5]
gopher://circumlunar.space:70/0/~solderpunk/phlog/two-walls-good-four-walls-bad.txt
[6]
gopher://sdf.org:70/0/users/tfurrows/phlog/2018/ah9_diskEncryption.txt
