151 reputation
7
bio website
location
age 26
visits member for 3 years
seen Sep 15 at 21:32

Jul
6
comment How does the “Fourth Dimension” work with arrays?
@cHao My remark on dimensionality in Haskell was exactly about types.
Jul
6
comment How does the “Fourth Dimension” work with arrays?
@cHao: That's undoubtedly how Martin Richards got his ideas for BCPL's memory model too. But don't forget that formal descriptions of algorithms aren't limited by what we can build, because they are inherently distinct: what we can build, like CPUs, is physical, while formal descriptions, like arrays, are ideal. For instance, using the description [[Integer]] (a list of lists of integers) in Haskell as a two-dimensional 'array' is inherently two-dimensional, even though it's all thunks and pointers when you compile it for a physical computer.
Jul
6
comment How does the “Fourth Dimension” work with arrays?
@cHao: implementationwise, there is space, because there is also time. The term 'space leak' (as an alternative for 'memory leak', found in the Haskell community) is no coincidence. The fact that memory is described as one-dimensional is a heritage from BCPL.
Jul
6
comment How does the “Fourth Dimension” work with arrays?
@cHao Maybe if you look at the semantics of the data they store. But on the representational/syntactic/implementation side of things, all array dimensions are inherently spatial. It's actually what you depend on when using arrays as part of an algorithm.
Jul
5
comment How does the “Fourth Dimension” work with arrays?
"Not all dimensions are spatial." For arrays, all dimensions are spatial.
May
20
comment What makes a language Turing-complete?
I wanted an authoritative source, and that Wikipedia article doesn't cite any (nor is it one). I don't really believe a word of that article, by the way. I'm currently writing an essay titled "Is C Turing-complete?", digging deep into the underlying CS and its history, and I can tell you with relative certainty that a claim like "C is Turing-complete" is nonsense; nevertheless, the article simply claims it, without any cited proof.
May
20
comment What makes a language Turing-complete?
Anyway, I interpret 'Turing-complete' as being simulation equivalent to a Universal Turing Machine (UTM; which, in turn, is capable of simulating any Turing machine -- hence 'universal'). In Turing's paper from 1936, where he introduced his machines, he defined the notion of a UTM, and gave a sketch of a proof that UTMs are simulation equivalent to Church's lambda calculus. By doing so, he proved that they had the same computational power. The Church-Turing thesis asserts, put simply, that "that's all the computational power you'll ever get".
May
20
comment What makes a language Turing-complete?
I'm still looking for an authoritative definition of the terms Turing-equivalent and Turing-complete (and Turing-powerful). I've already seen too many cases, from people on message boards to researchers in their own friggin' papers, who interpret these terms differently.
May
20
comment What makes a language Turing-complete?
"A programming language is turing complete if you can do any calculation with it." That's the Church-Turing thesis, not what makes a language Turing-complete.
Feb
6
comment Anonymous software license
That's an interesting point, but putting something that looks like a real name on top of your code isn't real proof of accountability (in case of stolen code). If there's anything that's easy to do these days, then it's creating a new identity. If you want it to look legit, you could even fabricate a life using social media. When I look at an AUTHORS file of any interesting project, I've never heard of most of them. How am I supposed to know to which natural person to turn if someone sues me for intellectual property theft, if those names might as well be fake?
Feb
6
comment Anonymous software license
But the burden of proof is with the claimants, right? What if I made it impossible for anyone to make a probable claim to authorship? That would of course be more than half of the work, but I think enforcing that property can be partially automated (don't publish it if natural language analysis shows that the documentation is distinguishable from a large English language corpus; do something similar for code).
Feb
6
comment Anonymous software license
@TRiG: I wasn't aware of the CC0 license yet. Although the CC licenses are much too wordy for my taste, it's certainly an interesting candidate.
Nov
21
comment How should a developer reject impossible requirements?
Also mention that the requirements are lacking, because a person of old age would have quite a thumb, but would still need to have a simpler and clearer UI.
Nov
19
comment Any good reason to open files in text mode?
Interesting (it would be much more useful), but very language-specific. For most operating systems, this conversion step is probably implemented in Ruby itself, while newline conversion can be done by 'deeper' parts of the host OS.
Nov
19
comment Examples of limitations in IT due to different bit length by design
In IBM PC-compatible master boot records, the partition table can't express partitions that are larger than or start further than 2 TB. This is because the sector size is fixed to 512 (the first sector of a disk contains the MBR; as a consequence, you initially have only 440 bytes of boot code) and the offset and size can at most be 2^32-1 sectors. To fix this, GUID partition tables are used now.
Nov
19
comment Any good reason to open files in text mode?
That sounds arcane. Mainframe-like arcane. Could you name the operating systems in the first two cases?
Nov
19
comment Any good reason to open files in text mode?
I get easily confused between SO and programmers. My apologies. If this is out of scope for p.SE, can it be moved to SO instead? My reason for posting it here, is that it's more about general guidelines than specific programming questions. I'll update the post for context as well.