Ask General Programming Questions Not Worth Their Own Thread

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > The Cavern of COBOL > Ask General Programming Questions Not Worth Their Own Thread

PhantomOfTheCopier: Aug 13, 2008; Pikabooze!

KillHour posted:

...
The graph is made up of "triples" which are a set (in the mathematical sense) of facts of the form [subject] [predicate] [object], where subject and predicate are both IRIs* and object is either an IRI or a literal.

So the smallest piece of data you could have is something like
code:
<www.example.com/obj#student1> <www.example.com/def#attends> <www.example.com/obj#school1>

Caught up to this point and am finally ready to ask: Have you looked at Prolog? It came to mind fairly quickly but it's still not clear what breadth of behaviors you hope to support in your "language that actually does or doesn't permit this distinction in a substantively non-pointer fashion". Since a logic language focuses on predicates, rules, and atoms, it seems a natural choice at least as a consideration.

Ultimately though you'll end up back at realizing that SQL or similar doesn't model your question painlessly, and invariably what's provided for business use needs to be at a different layer, and hence a different language: Almost always non-technical, unoptimized, and quite limited.

(Sorry, your discussion was 1mo ago but on the previous page so yeah justpost shrug.)

# ¿ Dec 1, 2023 04:51

Adbot: ADBOT LOVES YOU

# ¿ May 14, 2024 09:28

PhantomOfTheCopier: Aug 13, 2008; Pikabooze!

Are these the same boomers who are finally learning to never click on random links in email because 99.99% of the time it's a phishing attack? Unless you send them to a real website it seems like engagement could suffer, especially if most of the audience has been using MSOffice for decades and is scared of "all that cloud stuff".

# ¿ Jan 6, 2024 08:07

PhantomOfTheCopier: Aug 13, 2008; Pikabooze!

The search terms you seek are canonical representation or form, which is often used as a state space representation.

At a very high level it could be as straightforward as "the lexicographic key ordered json output of all (basic) variables". For games this is almost always position, orientation, speed, and other physical parameters. Objects, scoring, historicals, may also matter. Sometimes this is inherently lossy. For example, it may be sufficient to start in the correct "area" even if the position is not equivalent.

A more efficient canonical form generally maps the bits of information, as nielsm showed above, but this requires a binary channel of communication. Encoding is when you take your binary state and perform a conversion to avoid protocol issues, eg Base64. Generally canonical forms will provide a binary conversion, because it's easy to analyze mathematically, but you can also have direct conversion to other encodings as needed.

Developers love to get clever here, so you may see short lowercase alphas (26^6=308M versus 64^6=68B), or just alphas (52^6=19B) which users will likely get wrong with manual re-entry. (The worst is when teams decide they need full UUIDs, which are never "human readable", even though they only need a few bytes of entropy and don't have distributed concurrency concerns anyway). If entry is entirely manual there are options like word-based encoding (horse stapler) as sometimes used in s/key implementations, or now in what3words. Users can likely handle some numbers, however, so "Eddy tide comb 7663" can be adequate, but you'll still be limited to smaller spaces.

Also consider what needs to be plain/visible, obfuscated, or encrypted. IE if data is already known to the user and doesn't impact application behavior, it doesn't need to as guarded (eg username). cf check algorithm (crc).

# ¿ Jan 14, 2024 15:34

PhantomOfTheCopier: Aug 13, 2008; Pikabooze!

640 should be enough for anybody!

# ¿ Feb 29, 2024 04:20

PhantomOfTheCopier: Aug 13, 2008; Pikabooze!

Polio Vax Scene posted:

Which programming language is most closely associated with satan?

So stick with me here, Satan is typically seen as causing people to move toward sin, but all evidence suggests that people are abject sinners and those who most profess godliness/absence of sin and horror at it (note I didn't say 'enlightenment') are those most frequently found sinning. Ergo, popularity and claims of superiority are the marks of the sinners.

There are only two choices: Java and/or Python.

It seems likely that strongly and weakly typed languages are equally leveraged by the Morningstar.

# ¿ Mar 1, 2024 23:30

PhantomOfTheCopier: Aug 13, 2008; Pikabooze!

Reminder with the rectangle/triangle question, there's a difference between how many requests there are to perform a drawing action versus how many shapes appear on the result. In this case, all rectangle drawing commands fill the same region, so you can call f() a trillion times and you'll still only have a single rectangle. (But as written you have zero until g is called).

But triangles? Always zero since a triangle with base and height 100 necessarily has an irrational hypotenuse, which can't be drawn on a digitized medium. :11tea:

(cf the student thread, since such a response will drive any instructor into an uncontrollable rage)

Does the question literally say "0, 1, 2, or 3"? They may be expecting "0 if g is never called, 1 for g(orange,2), 2 for ..."

# ¿ Mar 14, 2024 01:54

PhantomOfTheCopier: Aug 13, 2008; Pikabooze!

Research is actually good stuff, very exciting, and wow I miss it, but really it mostly disappeared in 2005ish and has never recovered. Oh sure sure there are "new things", but in the world of theory we're still implementing a lot of things from the 50s and 60s.

Research is the first thing to get blamed/cut when revenue drops 0.01%, so it's something of a thankless endeavor in an industry where "planning ahead" reaches out to the 9mo mark, which is about the point that the fastest research starts showing results.

# ¿ Mar 19, 2024 00:33

PhantomOfTheCopier: Aug 13, 2008; Pikabooze!

tldr: Prepare to build something new and disconnected. Go read The Atomic Chef first.

M_Gargantua posted:

But this is plenty for me to research so I can better stress to program management the importance and ease of doing this right.

Those are the correct technical concerns but you'll want to go into the conversation stressing the actual goals, the need for some solution, and that the request is to understand the options and tradeoffs. While it's not particularly aggressive, "doing this right" could be interpreted, particularly by engineering, as "the solution has been decided, make it happen or be fired".

Control software... Let us assume the worst scenario: Someone forgets a preprocessor directive and control rod insertion fails, hence meltdown. Or MAX software lets the plane crash. Depending on the safety requirements, it may be useful to start with the worst risk and assess likelihood with respect to the proposed changes. EG "consultants hired for the simulator create a new flag that the control engineers aren't informed about, hence the production system fails"; IE the safety model gets broken because an external factor introduces an unqualified process.

There may be dozens of engineering reasons that prevent the above ideas from being viable options. Of course we'd like modular software, interchangeable libraries for different environments, but plan to go into the conversation listening intently for any concerns. Embedded systems, control systems, may all be specialized to the point where they won't support that (eg having enough memory or storage for the additional object references). EG, ponder the stories you've heard about the computers on the space shuttle.

There's a good chance you'll never touch the control software in this project. There may be very good reasons this should only ever be a simulation. Maybe it can be fed data from real system metrics or reports, but bidirectionality might well be more trouble than it's worth.

ps https://en.m.wikipedia.org/wiki/Knight_Capital_Group#2012_stock_trading_disruption

Flag-based software goof.

PhantomOfTheCopier fucked around with this message at 13:26 on Mar 23, 2024

# ¿ Mar 23, 2024 12:51

PhantomOfTheCopier: Aug 13, 2008; Pikabooze!

Ah always the most fun teaching Linux courses or pair debugging with people who spend their lives clicking little buttons with a mouse who need to do something from a shell: "Go to /etc"

# ¿ Apr 7, 2024 02:12

PhantomOfTheCopier: Aug 13, 2008; Pikabooze!

Seems like ndx would be the index associated with the data. https://fileinfo.com/extension/ndx

This matches with the lbl being a dbase file, though the .cal may still be readable with existing calendar tools. The ovr was likely Windows proprietary, maybe an SVG like label/drawing format.

Fortunately they are unlikely to be encrypted, possibly some minor compression, probably fairly easy to reverse engineer.

# ¿ Apr 12, 2024 00:03

PhantomOfTheCopier: Aug 13, 2008; Pikabooze!

Typographically, "insufficient leading" (that being pronounced "led", as in Pb) or "excess kerning". Graphically it would translate to an improper/insufficient bounding box, because the (likely vector) ascender position has been miscomputed when the italics/slant variant is in use. Without additional data it's hard to say if it's clipping or overwriting by the subsequent character (though I'd tend to agree with the others that it's being clipped to its bounding box).

# ¿ Apr 13, 2024 06:18

PhantomOfTheCopier: Aug 13, 2008; Pikabooze!

A Festivus Miracle posted:

pre:

number = int( input ("Enter a value greater than 0"))
product = 0
while (number >=0):
     for x in range (1, 3):
          product = number * x + product
     number = int( input ("Enter a value greater than 0"))
print (product)

I have no loving clue what it does, but whatever it is it does it wrong :v:

No, this may be the first thing actually correct: If I remember my python and this is one of those very awful range operators that actually means "from one to two spelled like 3", then the inner loop is always product+=(2+1)*number, so it just returns the sum of the entered numbers multiplied by three; and entry is terminated via a negative number.

IE it's a simple example of obfuscated code.

A Festivus Miracle posted:

she's absolutely stuck on dogmatic ideas of what code should look like.

The only thing I can say "to be fair" here is that it appears to be a lot of Python which is itself hopelessly dogmatic about what code should look like. As the student you may be unqualified to evaluate, but if true, then yes you've basically found the worst type of career programmer, one who can't read code and indeed refuses to do so unless all submissions are exact duplicates of what they would have written.

My advice, write your code, use the pep8 tool if it doesn't completely destroy your code, then find an external/separate tool that can handle custom formatting options and use that to get submissions closer to DummTeacherCode.py. Advanced level is to write the tool yourself, put it on GitHub maybe, but certainly use it as interview leverage. "The simplest way is to use tools to achieve some code parity so teams can spend their time arguing over things that actually matter... like variable names :buddy:

PhantomOfTheCopier fucked around with this message at 02:43 on Apr 26, 2024

# ¿ Apr 26, 2024 02:29

Adbot: ADBOT LOVES YOU

# ¿ May 14, 2024 09:28

PhantomOfTheCopier: Aug 13, 2008; Pikabooze!

It's also a topic that nearly everyone will get wrong, because they'll approach it from their very narrow experience instead of invoking linguistics or psychology. The easiest way to see this is to pick what someone believes is an immutable rule of style then ask them to come up with all the exceptions to that rule. Justifications are likely to include "People do it that way" and "I think it's easier to read", but very little will be quantitative or even measurable.

Another fun exercise is to take someone's three "most important" rules and find contradictions.

# ¿ May 14, 2024 07:17

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > The Cavern of COBOL > Ask General Programming Questions Not Worth Their Own Thread