Discussion about Dyalog's integrity

ngn

14:24

@Adám after your disgraceful actions in the 2019 competition, i don't feel like competing

pxeger

do I want to ask?

ngn

@pxeger was that addressed to me?

pxeger

@ngn you and Adám - do I want to ask for the context to that remark, or will it stir drama?

Adám

@pxeger It is probably off-topic anyway, but I'll not stand in the way of ngn airing grievances.

Of course, it just may be that they are based on misunderstandings or misinterpretations.

Some hurtful actions is not caused by malice.

@ngn Is that's why you didn't participate in 2020 either?

ngn

@pxeger the drama is over. i have strong reasons to believe that in that competition they deliberately prevented me from winning. adam rudely refused (on behalf of dyalog) to publish the winning solutions and accused me of "personal" attacks against him here in stackexchange's chats.

pxeger

14:30

ok, let's not continue this discussion then

unless you particularly want to, in which case we should go to a new room

ngn

@pxeger that's all there is to it, there's nothing to discuss

Adám

For the record, we didn't want to publish people's submissions without their explicit approval. Based on ngn's feedback, we've begun asking for such approval this year.

Adám

Yeah, let's let it rest.

ngn

@Adám don't tell people what to do

Adám

@ngn Oh no, it was just a proposal. By all means, go ahead if you want to go over it!

rak1507

14:43

you didn't publish solutions this year

despite saying you'd ask for feedback to

and also changed the judging method

so I'm not really convinced that it's becoming better

Adám

@rak1507 Uh, we're still planning to do so.

rak1507

how long does it take to upload a text file?

Adám

@rak1507 Wait, were we not allowed to do so?

@rak1507 I already said that we're gathering permissions.

rak1507

@Adám of course you're 'allowed', it would just be nice to know how things work rather than putting your code through the opaque dyalog machine and hoping for the best

Adám

> Phase 2 will mainly be judged based on:

∘ Did you solve the problem?
∘ Does your solution demonstrate appropriate use of array-oriented techniques? Solutions that use looping where an obvious array-based solution exists will be judged lower.
∘ Did you comment your solution? It's not necessary to write a novel, or add a comment to every line, but comments describing non-trivial lines of code are advised. These help the judging committee determine your level of understanding of the problem and its solution.

So, first we checked for correctness based on a bunch of peer-reviewed test cases.

Then a group of experienced APLers looked at the most correct solutions to determine their use of APL (i.e. array-oriented techniques), and how clear they were (comments, naming, code style). Finally, we disqualified any submissions shared by multiple participants.

rak1507

14:52

but how can participants verify that if you don't post the ranking

for all we know you could have just put it through a random number generator

and I hardly trust the dyalog test cases considering the phase 1 problem 7 (and the fact that the language itself is full of bugs)

Adám

@rak1507 Huh, we do publish who got the three prices. Isn't that a ranking? And if we get permission, we'll publish their code too, so you can verify (to some extent, because some of the criteria are subjective) that those submissions are indeed good.

rak1507

Weren't there 4 prizes? Not sure how that 4th one was invented but it was apparently.

Adám

@rak1507 Ah, yes, the professional prize?

rak1507

No, the 'special recognition'

Adám

That participant did not receive any prize.

rak1507

14:56

Phase II – special recognition

    Mehdi Zribi (Liceo Scientifico GB Grassi Saronno, Italy)

receives a cash prize of $500 USD

Are you sure?

Adám

Wait, what‽

I was unaware of that.

rak1507

Certainly fills me with confidence when a dyalog employee is unaware of large sums of money being given to people without their knowledge :P

Adám

While I'm not in charge of anything money related at Dyalog, I'm certainly going to enquire about that.

rak1507

What is special recognition anyway, what was so special? I saw their solutions on github, looked like perfectly normal average solutions to me.

Adám

@rak1507 Well, I thought it was just a recognition of their solutions being good, even though they didn't win. But now that there's money involved, I don't know either. Rest assured that I will investigate. (However, getting answers on a Sunday might prove hard.)

rak1507

15:02

So if your solutions are good and Dyalog feel bad for you, you get given a prize anyway? Damn, the olympics should do that, just give everyone the gold!

Adám

So it does seems at the moment, but this is rather useless speculation.

But all this said, the competition is almost free to participate in (other than time and computing resources the participant uses), with substantial prizes being given out. I really don't think anyone is entitled to complain.

rak1507

Yeah true

Adám

(You probably meant that as an answer to my first message.)

rak1507

Yes, but it also applies to the second a bit, although I think it would still be better both for Dyalog and the participants to make the competition more open etc

ngn

complaining is almost free too, so why not :)

Adám

15:06

@rak1507 I suggest you write an email to Dyalog sketching up how the competition could be made more open etc. If formulated sensibly, then I promise you that it will be read, considered, and replied to.

pxeger

@Adám so far it seems to have caused some discontent in the APL community, which I imagine is not the goal of these outreach events. So it's probably in your interests to entitle people to complain

rak1507

@Adám will do, I'll do a bit of research into if there are any competitions that judge the code quality too as well as the correctness

Adám

@pxeger Yeah, that's a bit sad. But constructive criticism is fine (and welcome). Many (if not all) of the complaints about Dyalog do appear to be solely for the purpose of stirring up matters, not actually because of anything substantial.

rak1507

That's the kind of comment that could cause people to get even more annoyed

ngn

@Adám reeeally??

rak1507

15:10

see ^

ngn

lol :)

Adám

@rak1507 Right, please do. It is a significant work on our part to do so, and we do that to encourage quality submissions, so people will have an incentive to become good APLers. It isn't just for entertainment. Indeed, some competition winners are offered jobs by our customers.

rak1507

Yeah, it's an interesting thing, as far as I know there aren't any others that do, but there probably are. It adds an interesting component of subjectivity, but I think the objective things (does the code actually work) are still a bit obscured relative to other contests

Adám

Maybe we should publish our test cases after (because finding out what various cases and edge cases are, is part of the challenge) the competition. That'd at least make the objective parts more open. Feel free to suggest that.

rak1507

Yeah

ngn

15:15

@Adám yeah, and do publish a cryptographic hash of the tests before it

Adám

@ngn Why?

ngn

@Adám because your behaviour doesn't inspire trust

Adám

@ngn Do you trust anyone?

ngn

@Adám yes

Adám

@ngn Yourself and dead people?

@ngn What would the issue be with adding additional tests after the competition has launched, if we're not applying those tests until after the competition has closed?

ngn

15:18

i'll leave those unanswered

rak1507

Yea it seems fine to me to just publish the tests used for judging, I don't believe dyalog would deliberately leave out/include test cases to benefit specific people

As long as the test cases can be reasonably judged as 'comprehensive'

Adám

@ngn OK, but then your claims are without basis. Why should we trust you? Have you shown a consistent record of verifiable trustworthiness?

ngn

@Adám i'm not organising a competition. you are.

Adám

True and irrelevant.

ngn

and no, you haven't shown a good record of trustworthiness

Adám

15:20

That wasn't the question. But that claim is still without basis.

You're making unfounded claims, which could have merit if there was basis for trusting you. But as things stand, it looks like you're simply out to slander.

ngn

cmp: should i flag ^this?

rak1507

No

user

I don't know about that, but I would highly suggest moving this entire conversation to another room

pxeger

I would like someone to move it all to a new room though

ninja'd

PyGamer0

@user which room though??

pxeger

15:27

create a new one

user

A new room just for them to argue or whatever it is they want to do

Adám

I'm making one.

ngn

i don't want to argue

Adám

← 90 messages moved from The Nineteenth Byte

pxeger

@ngn This is sounding a lot like Monty Python's Argument Clinic

Transcript for

Discussion about Dyalog's integrity