The Whiteboard

12:00 AM

@Joe Requesting close votes. You don't have enough rep.

@MattD Then why use the old? It sometimes said that MS doesn't support new technologies forever, but it's unquestioned that when they phase something out they stop supporting it.

Matt D

just wondering about people whove actually used both and opinions

im going with the new

Jimmy Hoffa

Ah, well I've not used the new one so I can't say. I'd be hard pressed to believe it isn't just everything the old one had but more and better considering membership was well regarded to begin with.

Joe

@RobertHarvey I'm happy with getting over 200 today ;) I can now comment everywhere, which is what I needed.

user20683

@ChrisCirefice Perl

Robert Harvey

12:02 AM

@Joe You could do that at 50. :)

Joe

@RobertHarvey Somebody already mentioned that before. You need 200 to get the +100 bonus on other sites, which allows you to comment on those, which is what I needed.

Jimmy Hoffa

@WorldEngineer I was thinking the same thing just because of Wall but I don't know that I agree completely, granted I don't know Perl particularly.

Based on his experience I would say JavaScript just because frankly, unless this dude's some MIT genius this project is one damn lofty one, throwing a new language in the mix doesn't really make sense if he just wants to make something kind of work.

Joe

It's tough to define what is a good question on OnStartUps.

Robert Harvey

0

Q: Programming Languages Best-Suited for Language Analysis

Background: I am about to go to France (in January). There, I'll be studying French linguistics; more specifically grammar and phonetics. While I'm there, as a learning exercise for both my majors (Computer Science and French), I will be attempting (could fail miserably) to design a lexical analy...

Too bad it appears to be off-topic... it's a well-written question, if a bit broad.

user20683

@RobertHarvey Invited to chat.

user20683

12:09 AM

pity MichaelT isn't here.

user20683

@JimmyHoffa I'd argue maybe an ML. Ruby's got good Regex.

Jimmy Hoffa

@WorldEngineer I'd handily argue Haskell but see my second point... all his experience is imperative, throwing him into real FP would be a lot

Functional languages tend to be better fit for analytical purposes IMHO, but that's just an opinion and those are all you can get to your third question there, so I'm goign to remove that one if you don't mind terrible as it really can't be answered authoritatively, it's a gorilla vs. shark question. — Jimmy Hoffa 1 min ago

I wasn't sure if off-topic or not

I just clipped out the 3rd question, the second one is almost a list question? So maybe you're right, because the first question hardly has any substance..

user20683

@JimmyHoffa Scala maybe

Robert Harvey

0

Q: I'm currently behind two NAT's with no public IP. Is there a way to self-host a client/server application?

I'm planning to write some small application for me & my friends. Specific functionality aside, it should allow me to "host" a session (using a dedicated server application) and my friends to join using client apps. I'll be using .NET because I like it the most. I'm a bit worried about connectiv...

c# .net wcf server

Decent question, but maybe it belongs somewhere else?

Jimmy Hoffa

@RobertHarvey "What technology to write a client/server application in?" -> Too Broad.

> Technical side - WCF all around? And if yes - what binding? If not - what instead?

Robert Harvey

12:14 AM

@JimmyHoffa He doesn't want the whole Bible; he just wants to be pointed to the right chapter. Sockets, WCF, etc. Assuming, of course, that it's even possible at all.

Jimmy Hoffa

that part sealed it for me. Summing that up as a little by-note when there's an enormous amount that goes into picking a networking stack and it's configuration details for any given purpose... We could play 20 questions with him all day to find out what would work there

Robert Harvey

What would work there?

Given that there's no public IP.

Jimmy Hoffa

@RobertHarvey Multiple possibilities, which is the problem.

@RobertHarvey The fact that it's natted is really a networking detail that could be unwound in the networking stack between the external and internal for him, make his router behave properly and his internal stack wouldn't know the difference.

Chris Cirefice

¡Hola! I see my question was closed ;) for valid reasons I suppose :) thanks for the direction nonetheless :)

Robert Harvey

Welcome to The Whiteboard. Where ghosts abound.

Jimmy Hoffa

12:18 AM

@ChrisCirefice Go learn Haskell, french isn't even a programming language booriing

Robert Harvey

(Dammit, I knew I left that ammonia-based cleaner somewhere.)

@JimmyHoffa Now that's good advice. :/

Chris Cirefice

Thanks Robert! Haskell, sheesh. Someone in my Structures of Programming Languages class is learning that, he says it's a nightmare :P

Jimmy Hoffa

kidding aside, for lexical analysis FP languages have tons of tooling and focus on that because much of them are written by and for language researchers (programming language researchers but still they have to do lexical analysis)

Chris Cirefice

That is kind of what I figured -

Robert Harvey

It's a recursive nightmare. It's H A S K C E P T I O N

Chris Cirefice

12:20 AM

I was leaning toward Ruby honestly, because that one feels most at home with me

user20683

@RobertHarvey Mal is what happens when you have a monad error.

Jimmy Hoffa

Nov 6 '12 at 20:34, by Jimmy Hoffa

The crazy thing about haskell is the levels. After you get over the first hump, I literally couldn't sleep for most of a week when I first started learning haskell because everytime I closed my eyes I had my brain kept trying to solve problems using syntax it didn't even understand. Dreaming in code sucks, dreaming in a code you have no comprehension of at all is even worse

Chris Cirefice

HA

user20683

@ChrisCirefice You can always refactor it.

Robert Harvey

@JimmyHoffa That was a year ago. Has your brain healed yet?

user20683

12:20 AM

Prolog and its relatives could be useful as well.

user20683

@RobertHarvey No. We all know that.

Robert Harvey

Not necessarily referring to the alcohol consumption.

Jimmy Hoffa

Mmm black grouse at home...

Chris Cirefice

Hm... I was hoping to take an object-oriented approach to the problem - building parts of speech, giving them their own classes and methods to check for certain context elements that would give better insight into the semantic meaning of a "sentence"

@WorldEngineer Prolog, I don't know that sounds hard :P

But I do see that Prolog is used for computational linguistics, go figure!

Now, I have no experience with mixing languages and environments - is it possible to write certain parts of the software in one language, and another part in Prolog, for example?

user20683

yes

Jimmy Hoffa

12:23 AM

If you want easy you're in over your head, language analysis is hard

user20683

PhD level non-trivial

Chris Cirefice

Ha, this I know! Which is why I'm doing the basics - nothing too complex. A glorified spell-checker, if you will.

Robert Harvey

Any decent language ought to suffice, then.

Chris Cirefice

Haha, maybe :P I'm definitely interested in Prolog though, reading about it now;

user20683

@ChrisCirefice It's possible to essentially do functional programming in Ruby

Chris Cirefice

12:26 AM

@WorldEngineer Yes, I have written a few applications in Ruby that do matrix processing, primarily. I'm wondering though - which would be more logically complex to express the root of the problem?

Jimmy Hoffa

I think JavaScript's phaux homoiconicity makes it feel cleaner for FP

Chris Cirefice

Tokenizing words, building semantic meaning. I'm thinking about how I would write it in Ruby and it just seems like a bunch of Class inheritance with sloppy methods that relate them, as parts of speech

user20683

@ChrisCirefice I generally find it simpler to think in functional terms.

Chris Cirefice

@wo

user20683

In Haskell you'd do it as part of the type system

user20683

12:27 AM

in Ruby you'd do it as mixins I'd think

user20683

each type of word has a mixin that defines how it can be used

Chris Cirefice

@WorldEngineer Yes, I agree - which is why Prolog looks appropriate. I also have a lot of experience in Javascript. In Ruby, yes mixins, in Javascript through prototype chains :P

Which seems to abstract the problem appropriately I think :)

It's just that this is going to be a year+ long project, something that I'll be working on for a long time. I just want to make sure that I choose a language which can suitably express, well, a language!

user20683

yieldprolog.sourceforge.net

user20683

Yield Prolog

Chris Cirefice

Personally I wanted to build this in C# as yet another learning exercise - interesting that there is a Project that allows you to use Prolog in C#!

Honestly though, I have no experience with mixing languages and struggle to understand how it's even possible, let alone how I would do it.

Jimmy Hoffa

12:31 AM

Blech closure composition not prototypes, inheritance is so the wrong tool

user20683

@ChrisCirefice If you've got any Java experience, C# will feel very familiar.

Jimmy Hoffa

For this job

Matt D

inheritance is often the wrong tool :0

;)

user20683

Inheritance almost always seems like the wrong tool but that's probably the Haskell talking

Chris Cirefice

@JimmyHoffa I'm sorry, I guess I don't understand that.

@WorldEngineer I do have Java experience, and I have read a lot of C# code. It just seems to me that there is a lot of "bulky" syntax, to declare Classes and inheritance and all that.

user20683

12:33 AM

@ChrisCirefice There is. Boilerplate is the name of the game with those two.

Matt D

we're suggesting that you dont use inheritance unless you absoluetly have to

Chris Cirefice

@WorldEngineer Yep, that's why I wanted to avoid them - but at the same time, the syntax used itself is "plain English" (in my eyes) as to what the whole system does. And specifically for languages, it's appropriate I think

@MattD I would agree - which is why Ruby looks nice for this project. I could have parts of speech as classes, representations with their own methods for determining which other language element surrounds them, and build semantic meaning from there

user20683

ohloh.net/p/ruby-linguistics

user20683

ohloh.net/p/nltk

user20683

Python version is massively more developed (because Python has somehow become this thing for SciComp)

Chris Cirefice

12:36 AM

Wow! That Ruby Linguistics project definitely looks interesting!

Matt D

for language. i'd probably have a very shallow heirarchy of different types of language "parts"

Chris Cirefice

However, as a caveat, the point of this project was to demonstrate my own understanding of French linguistics as well as software design ;)

Matt D

and then compose sentances and what not using those parts

the thing which composoes sentances understands the rules of the language

Chris Cirefice

@MattD That's exactly what I was going to do :)

Matt D

and what things can be plugged into other things

(and their order)

and how the order may change the represntation of the element which is composed within it (changing sex and what not)

Chris Cirefice

12:37 AM

Just wanted to try to find a language that would let me do that appropriately without the overhead of convolution

@MattD That's basically what I was going for, aside from the spell-checker aspect of it ;)

Matt D

you could also use a similar system to convert between languages

which i guess is how those systems work anyway

Chris Cirefice

Well I DID want to learn Russian... ha

Matt D

well, you'll have nouns, verbs, etc etc, if you had a list of syntax "rules" the component pieces will be roughly similar

assuming you can map lexicology

which given the output of google translate is probably exactly how they do it.

yeah. that makes sense. and you could use the same rules to assert grammattical "checking"

sounds like a fun project

Chris Cirefice

Mhm! There are somewhere around, oh I don't know, over 9,000 grammar rules in French, with plenty of exceptions. My goal was to at least catch half of them by the time I came back to the US ;)

Matt D

rules are just data

if you can get a list of rules, and then load them as data....

Chris Cirefice

12:43 AM

Yep, that's true. Functionalism is what would allow me to evaluate the rules, but I would need to build data structures to be evaluated in those rules

I think that Ruby is probably the best choice -

Again, having no experience with mixing language, is there a way I would write part of it in Prolog as well?

Even if not practical, it would be a good learning exercise in any case. And learning Prolog, why not!

psr

@MichaelT oh the irony!

user55340

@psr Its one I wish I stuck with, though it wasn't the "how do you do sql" its was "how do you write the code to do various types of indexing and transaction isolation and such"

Chris Cirefice

Well, I need to get going home, just got off work. Thank you all for the suggestions, it is very much appreciated :) And good night!

Jimmy Hoffa

1:21 AM

@RobertHarvey Aye, it took a lot of bashing my head with Haskell for the basics to fit in, then I had to start hitting it harder for longer periods to get the fundamentals, and I'll let you know when I've got those in there. The basics were just like a headache for two straight weeks, but after the headache subsided I could do things like whip up that snippet of C# the other day just off the top of my head. The wikipedia stuff I linked earlier is fun interesting thought material;

but Haskell itself is definitely more than mental masturbation, no question about it: I can think through a variety of problems wayyyy easier and clearer than before I learned it.

jozefg

3:35 AM

@JimmyHoffa Well there you go, a nice, category-theoritic approach to STLC like type systems by forming a closed cartesian category

Jimmy Hoffa

4:17 AM

@jozefg Is my answer a bit reaching? I know it lacks a lot of formality, but just thought I'd try to draw his conclusion out best I could. Did I get that right in my comparison to lambda mu or is that more akin to STLC?

Does it seem at all accurate?

and I know about the out from termination but you lose totality checking in all those cases don't you? That's why I mentioned that in particular instead of just dependant typing overall

gnat

7:28 AM

@JimmyHoffa "+1 great query!" data.stackexchange.com/programmers/query/150346

gnat

7:46 AM

Matt D

7:56 AM

?

gnat

8:11 AM

15 hours ago, by MichaelT

That means... we need more comment flags!

gnat

12:00 PM

health tag looks... sick, could we use some surgery here please? 13 weird questions, 11 of which are closed

-1

Q: Does actively programming help with mental health decline?

Playing puzzles such as crosswords and Sudoku have been shown to reduce mental health decline, and computer games such as Lumosity claim the same thing. This has lead me to think: what about programming? It's an analytical task that isn't that far from solving puzzles. Has there been any researc...

health

enderland

12:59 PM

18

Q: Does solving Sudoku puzzles improve your mental abilities?

A lot of people I know see solving Sudoku puzzles as a way to train their minds. howtoincreasememorypower.com recommends doing a Sudoku puzzle a day to increase brain power. Researchers rank solving sudoku puzzles daily among the top ten non-traditional and alternative ways to boosts brain ...

iq

gnat

2:12 PM

@enderland I bet Skeptics wouldn't be able to match all the specter of healthy sickness here...

37

Q: Can passionate programmers with limited abilities have life outside work?

Update: Thanks for all the wonderful answers posted. It really helped me to put things in perspective. I definitely feel better now and signed up for gym few days back. Also I leave my work laptop at office, to avoid working late at nights:-) Anyway I started sleeping for 8 hrs and it feels good....

work-environment health

enderland

@gnat I'm actually worried about that. I like development... too much and it starts taking over my mind both on/off work :\

though aint no one got time for dat (coding 18 hours a day)

Transcript for