Comments for Software Sermon

Comment on Barbarians at the gates by cardiff space man

cardiff space man — Thu, 23 Mar 2023 10:49:49 +0000

In the beginning of the lock-down, I found that Eclipse CDT’s remote editing capability had gone a long time without updates. Some of my colleagues had jumped onto VSCode and so I decided to try it. Its remote editting works very well. Also I found that while Eclipse has a dialog box for adding a single entry to include paths, VSCode configures this with a JSON file and you can put in many entries at the same time. There are other differences relevant to C/C++ between the two. But these two made the biggest impression because I encountered them first.

Comment on Global Warming and Java by dayton@ieee.org

dayton@ieee.org — Fri, 01 Mar 2019 08:50:37 +0000

In reply to Dan. Help me out and point out the errors I made or the numbers I got wrong, or the bogus sources I used. I'm sorry that my sense of humor missed the mark with you. Of course a few per cent waste in power generation waste is nothing in comparison to the burning of rain forests, cement production, and fossil fuel power plants -- but every little bit helps. Do you really need to add to the problem with an inefficient program?

Comment on Global Warming and Java by Dan

Dan — Fri, 01 Mar 2019 07:11:01 +0000

Is this a joke? This must be a joke.

I don’t have a horse in this race. I am a mathematician and it needs to be said that this “evidence” is absolute nonsense. The idea that a choice of programming language contributes to global climate change in any significant way is ludicrous. (I won’t even get started on those laughable computations.) It’s as silly as saying that we should all stop eating Mexican food because there are greenhouse gasses in farts. (Yes, I intentionally picked an analogy that is both ridiculous and a tiny bit offensive.)

Programming languages are tools. Most of them excel in some applications while struggling in others and we all have our preferences. That’s fine. However, making ridiculous arguments like this in the face of serious discussions makes you look like either you’re an idiot or you’re so impossibly biased that you’ll latch on to anything, no matter how absurd, that pretends to support your belief.

I’m going to stop writing this comment now because I’ve run out of synonyms for “ridiculous.”

Comment on Offside! by cardiff space man

cardiff space man — Wed, 13 Sep 2017 17:08:55 +0000

I assume your diatribes against scripting and automatic memory management are in the queue?

I would like to point out that we may be up against it if we are trying to get rid of the offside rule. It is a very old rule and some of the coolest and most academic languages use it.

Peter Landin invented the offside rule in 1966, if I’m not mistaken, for the conceptual programming language ISWIM (https://en.wikipedia.org/wiki/ISWIM), which inspired all the well-known functional languages, and all of them adopted the rule, with slight variations. Of course ISWIM was described in Landin’s paper, “The Next 700 Programming Languages”. Haskell’s version of the offside rule amounts to (1) accept curly braces (2) use the offside rule to generate curly braces (3) if there are syntax errors at certain keywords then insert curly braces to see if that fixes the error. Scheme has SRFI 119 which proposes to adopt the rule into a LISP dialect.

Rust uses curly braces, and depending on the lifetimes you need for your state variables, doesn’t use a heap. Plus it’s compiled. Is it the ideal language?

Comment on Woodpecker Apocalypse by sdf

sdf — Mon, 28 Mar 2016 11:15:07 +0000

Aw, this was a really good post. Taking the time and actual
effort to generate a very good article… but what can I say…
I hesitate a lot and never manage to get anything done.

Comment on Woodpecker Apocalypse by cardiff space man

cardiff space man — Tue, 18 Aug 2015 01:10:13 +0000

Your sample code has boost dependencies but you said that C++11 has the features you need. Since I either write code for work or single-threaded code, I don’t know if you needed to use boost to fill in C++11’s gaps or what.

I agree with you that users of the double-checked lock idiom need to be corrected, but the reason should be given: The idiom doesn’t work if the code is optimized.

If you implement automatic unit tests the next thing that you should implement is getting the tests run with Jenkins (or a similar CI platform). After that, implement numerical requirements for test coverage and number of passing tests, automatically enforced with the help of Jenkins. I have heard of shops which enforce percentages of both in the 90’s.

And then you should implement automated functional testing.