The Whiteboard - 2013-07-15 (page 1 of 2)

« first day (1046 days earlier) ← previous day next day → last day (3950 days later) »

12:07 PM

Morning, programmers.

thorsten müller

12:42 PM

Hi, Boston HQ :)

1:11 PM

Morning!

user20683

1:29 PM

self.morning(sleepy);

while(sleepy)
{
drink(coffee);
}

also works for beer.

while(sober){ drink(beer) }

shit. forgot a semi-colon. Might as well go back to bed and try again tomorrow

user41796

drink(beer) isn't the worst method to be stuck in evaluation.

user41796

While reading over blogs & forum Q&A sites, I'm reminded just how much I love SE's format where the crap is actively culled out.

wait... are you saying that yahoo answers isn't a reliable site for finding answers to my every day questions?

user41796

No comment on any particular site. :-D But the signal to noise ratio outside of SE leaves a lot to be desired.

user20683

2:03 PM

@GlenH7 I ran for mod explicitly to keep it that way

user41796

2:29 PM

It's something that keeps coming up as I have other challenges in my little world to solve. Finding solutions to my lawn sprinkler system's leaks and digging up information for youth soccer has been .... challenging. Makes me happy to have my little corner of quality here on P.SE. :-)

Can I ask UML related questions on this site?

@cpx As long as it meets the other guidelines for what's on-topic, yes. Design questions (including about modeling languages) are on-topic.

I need to ask questions e.g "How to do this thing in Visual Paradigm for UML?"

That's on the fence. Tool questions are technically off-topic on Programmers and belong on Stack Overflow and it's defined that way in the FAQ. However, when it comes to tools that relate to things that are on-topic (design/modeling tools, for example), I'd like to see them on-topic. Not a lot of community support for that, though.

There are 26 questions tagged Visual Paradigm on Stack Overflow, though. And 45 questions mention Visual Paradigm‌, so there are people there who can help.

lawn.se and soccer.se done

user41796

2:43 PM

@JimmyHoffa close. sports.se & diy.se. Regrettably, no answers for what I was looking for. Both are beta sites and are a bit lacking at the moment.

user41796

And I'm surprised you didn't weigh in yet on my software reuse question

user55340

2:54 PM

@GlenH7 diy.se doesn't look too beta...

user55340

diy.stackexchange.com/questions/tagged/sprinkler-system

@MichaelT This. DIY is Home Improvement and it's def not Beta.

user41796

@MichaelT You're right, and my bad on calling it beta.

user55340

They've got 30-45 questions on sprinklers alone...

user55340

(and I can have more fun on that site asking questions and giving answers - was hesitant working for a home improvement retail company before)

user41796

3:05 PM

I think it was the lack of depth in the answers that I saw on DIY that pushed me into murkier sites. Which really means I need to just join DIY and get some answers in there. But I needed to fix my problems now

@GlenH7 Heh yeah, that's exactly the problem that has bad questions showing up all over SE, good on ya for knowing a need to fix your problem right now is not a good fit for asking a question on SE often times

user41796

"My sprinkler manifolds are leaking like a sieve. What should I do?" :-P

user55340

MichaelT,I m new to these sites. I'll take care of ur suggestion from now on. I don't think its off topic . SO ,SE are great websites most of the times i get my answers here. I created account recently so posted on multiple sites at the same time because it is very urgent for me to know the answer for my problem. I believe there are active expereinced programmers who help many people through these websites. — coder 8 hours ago

I'll occasionally take the time to construct a decent question for something I need right this minute, post it to SE and then just search out and about with all expectations I'll find a specific and correct fix myself before I ever get a quality answer from SE knowing the answer from SE will only be tangential to my actual problem

user41796

A: de-pressurize; clean the area; apply plastic epoxy stick (think JB Weld); allow that to cure; then wrap with self-sealing silicone tape.

user55340

3:09 PM

> because it is very urgent for me to know the answer for my problem

user55340

The other bit, chat isn't a bad place for "very urgent" questions.

Playing the devil's advocate, Chat doesn't see nearly as much traffic

user55340

Just have to acknowledge that we are not paid support for random software... unless you want to hire us as consultants.

That I can agree with

user55340

@Ampt depends on the chat.

user41796

3:10 PM

@MichaelT They might blink at my bill rate. :-)

user55340

@GlenH7 Same can be said of sprinkler repair people...

@GlenH7 A: sieve's are very efficient, I would suggest constructing another sieve near the leak that feeds the water back into the manifold for a recursive sieve correction which due to the efficiency of sieves will likely increase the pressure output of the overall manifold

...@MichaelT look, I could write a nonsense blog on any topic, it doesn't even have to be technical...

user41796

@MichaelT parts alone on rebuilding the manifolds would have been pretty expensive. Made me understand why commercial systems have one valve by itself and not use manifolds.

@GlenH7 as for why I haven't responded to your software reuse question, I pondered it for a bit and am pretty sure I don't have anything useful to say on the topic. In my book you're kind of on the money in pointing out a hole I've wondered about in the past. I chalk it up to "meh" because the fact is as the concept can be taken as a "We aren't doing this so we won't get the benefits from it" that's the totally wrong interpretation of the problem; the problem is:

"We aren't doing this, so we have no idea what would happen if we did"

Could be an improvement, could equally likely be a terrible effect on the result, second-system effect comes to mind immediately, but that doesn't completely apply

I chalk it up to one of those holes in the industry of which there are many where we lack sufficient study results because well, academics don't want to study boring crap like that, and industry doesn't study, industry works.

It's hard to construct study's for lots of things that are problems specific to industry, because without having an actual industry scenario you aren't going to replicate the exact dynamics, and businesses are too busy to help people out with studies or do them on their own

user41796

@JimmyHoffa Which is why I was hoping to see your counterpoint in there. :-) If you abstract it out, I'm asking if software development can become Engineering. I obviously scoped it done to avoid the challenges such a broad question brings. I liked Thomas' answer but was hoping to see a counterpoint in there.

user41796

3:19 PM

Patterns (sorry MichaelT) are a means to try and close that gap from the lack of repetition.

@GlenH7 Perhaps, do you call a plugin system a pattern?

user41796

One of my more provocative working titles was Does DRY keep software development from being real engineering

@GlenH7 Does real engineering try to create repetitions of a part of the system when it could allow a single instance to fill in?

user41796

nah, a plugin isn't a pattern. The term pattern has been abused. GoF was good starter material; and there was certainly more beyond the ones they listed. However, it quickly became a case where everything had to be a "pattern"

I think the problem comes from trying to compare software development to traditional engineering. The domains are inherently different, but I would argue that it takes just as much mental work and process refinement to do software development as it does typical engineering.

I also feel that we haven't really well defined the different types of people in the software development process

3:22 PM

Real engineering applies DRY too, it's just that they have less available scenarios where they can get away without creating a new piece

user55340

Excitement for the morning... catching a mouse... and not the computer type.

in building a building you have many groups of people, but almost everyone knows whoa foreman is, who the engineer is, and who the workers are

user41796

@JimmyHoffa Absolutely. Think of a bridge or a treatment plant. Those start from a basic pattern that covers all the things that have to be considered and have been optimized for efficiency. They are then modified specifically to the site where it will be constructed.

@MichaelT I thought your new place was supposed to be nicer... I guess maybe it is though, in your last place being part warehouse I know there's no way you'd catch the mice, they had entire forts surely

user55340

@JimmyHoffa They are repainting upstairs, so we got moved into the big (rarely used) conference room for the day.

user41796

3:24 PM

But the repetition of applying the pattern to the specific site is where additional efficiencies are discovered. There was work involved in the customization to the site. Not so with copy & paste.

@GlenH7 In real engineering, if an architect can come up with a way for one part to serve a single purpose to multiple other parts rather than having to be recreated so you have a separate part serving that purpose to each consuming part, he will do it. That's DRY.

user41796

@Ampt And the same is true with large scale software dev.

user55340

One guy was getting a network cable and disturbed a small mouse hiding in the podium... followed by catching it.

user41796

@MichaelT cage it and you've got a new team mascot.

@GlenH7 yeah, but it's not as widely known, and thus for a lot of smaller companies they give titles to people whom don't necessarily fill that role. This adds to the confusion of whether a software engineer is truly an engineer, because in many places the person filling that role is a software dev rather than an engineer.

3:25 PM

@GlenH7 If you didn't apply DRY to bridge building, there would be a separate anchor strung from the land for each individual pier in the golden gate bridge

user41796

@JimmyHoffa perhaps I erred in comparing bridges to authentication or logging systems then. I saw a bridge or a building as a broader system. Likewise with an auth or logging system.

user41796

Perhaps I shouldn't use DRY, as that's focuses within a particular project. I'm thinking about the feedback from designing / building the same class of system over and over. I don't think I'm calling out that broader aspect well enough.

@GlenH7 And this is why I don't fit "patterns" as the attempt to create repeatable parts, I think it more that we repeat concepts, plugin based architecture is a concept, REST is a concept, web service as a part of your program is a concept

@GlenH7 I understood what you were saying, but you served me up an easy problem to balk at, because the broader concept is really something no one can speak with any authority on, again because we lack relevant studies and details about what the results would be in either approach

heck, maybe I should close your Q as primarily opinion based for that alone :P

user41796

@JimmyHoffa Actually, I was worried about it being too broad and am pleasantly surprised to not see any close votes on it. and / or primarily opinion based.

It can be discussed, but no one knows one way or another with any detail exactly how it would play out

In my book it kind of falls back to the "coding is designing, not creating", and in that vein @GlenH7 I have been told in the past architects will frequently reuse parts of blue prints they've previously worked on

user41796

3:32 PM

If nothing else, I'll argue it's a welcome break from the ultra-localized gimme teh codez before this is closed questions.

When an architect has created a blue print for some building or another, and another project comes up that has one part that should be similar in part to one of the parts of their previous system, it is not uncommon for them to just take the exact same part from that previous one for the same reason we use 3rd party libs instead of writing our own: It's already been tested, weighed, measured, and shown capable

user41796

One of the benefits of working in an actual engineering firm is I get to see how the others branches of engineering put their work together.

(This is me speaking from anecdotal hearsay, I could be wrong, maybe they never reuse portions of blue prints)

user41796

@JimmyHoffa Oh, absolutely. And then they tweak it based upon the site it's going into. My conjecture is they are repeating themselves with the per-site tweaking. But now I'm just restating what I've already said.

user41796

No, I can confirm that is commonly done. CAD really simplifies that task.

3:35 PM

@GlenH7 When you use a 3rd party lib, do you not write an integration point in your own code where you wrap and deal with it for your systems particular purpose?

user55340

... another mouse.

You know what that means!!

This is especially true for Architectural Engineers and their cookie cutter buildings. Girlfriend does that doing HVAC and although every Gold's gym is pretty much the same, the location in the country changes what kind of HVAC they get. Lots and lots of tables to pull data from based on average temp/humidity etc

user41796

@MichaelT you'll want separate cages for your mascots. Just saying....

@MichaelT ThunderDome!! Two mice enter! One mouse leaves!

user41796

3:36 PM

@JimmyHoffa Regrettably, no. Most of the time, I see direct integration of 3rd party libs. Which creates a lot of problems later on down the line.

@GlenH7 Well now you know what happens when you don't custom fit the reused bits to your purpose :P

user41796

Maybe I should write my own countering answer.....

I always put even if it's a dumb pass-through, some integration point for 3rd party stuff, just because while I may not need it yet, in the future you never know when I'll need to have some interstitial step

(That's right, I used that word. You just got englished)

user41796

Interstitial? Common word for me. Then again, my background is in biomedical engineering. Interstitial tissue is a pretty common thing. :-D

plus the interface let's you mock it, as well as the fact that frankly it doesn't take a ton of work to implement compared to the time it saves

@GlenH7 You worked on cyberimplantronics? I love those!

user41796

3:40 PM

In grad school, yes, to a degree. Mostly basic research trying to map out neurons so you could build circuitry off the patterns that were firing. Wasn't the most rigorous lab though, so that work never progressed to the level it could have.

user41796

My thesis involved reworking and extending an alternate model that explained how neurons would fire.

That is genuinely awesome. Did you ever stab someone with one of those needle zappers for zapping their nerves?

user55340

@GlenH7 Ever read Accelerando? an AI built from lobster neurons is a character...

user41796

Like I said, the lab wasn't as rigorous or well funded as it could have been.

user41796

@MichaelT No, I haven't, but that sounds hilarious.

3:42 PM

@GlenH7 Well if you ever want to take it back up, @MichaelT has two mice you can start testing with.

user55340

antipope.org/charlie/blog-static/fiction/accelerando/…

He said they're not computer mice, but well, you could change that

user55340

Accelerando (novel)

Accelerando is a 2005 science fiction novel consisting of a series of interconnected short stories by British author Charles Stross. As well as normal hardback and paperback editions, it was released as a free e-book under the Creative Commons Attribution-NonCommercial-NoDerivatives license. Accelerando won the Locus Award in 2006, and was nominated for several other awards in 2005 and 2006, including the Hugo, Campbell, Clarke, and British Science Fiction Association Awards. Title In Italian, accelerando means "speeding up" and is used as a tempo marking in musical notation. In Stross' ...

user55340

Its about the singularity... and that link before the wiki link is to the author's site where the book is freely available.

user55340

One of the odd bits in there, they mention an idea from another author's SF. (This book is in turn mentioned in Implied Spaces)

user41796

3:45 PM

@JimmyHoffa - That's an area of research I don't want to get back into. I think simulation capability has dramatically improved and would be more interested in working on the tooling side than on the live study side.

user55340

It mentions a basilisk as being nothing more than a headache... ansible.co.uk/writing/c-b-faq.html

user55340

David Langford has a number of SF stories about basilisks.

user55340

Another author free (short story this time) on the basilisk - lightspeedmagazine.com/fiction/different-kinds-of-darkness

user41796

@MichaelT thanks for the links. Those are going to be added to my reading queue.

user55340

idea of the basilisk novels - there are certain fractals that if you see them, will kill you. Terrorists need only a can of spray paint and a stencil.

4:00 PM

@MichaelT Properly used a terrorist could kill you with a can of spray paint and a stencil as it is

user55340

@JimmyHoffa Thats one example... its an interesting set of short stories. I've read 4 of 'em, and linked two of the freely available author sanctioned versions here.

user55340

10. Is it true that Microsoft uses basilisk booby-traps to protect Windows Ultra from disassembly and pirating?

We could not possibly comment.

user55340

7. Is it just an urban legend that the first basilisk destroyed its creator?

Almost everything about the incident at the Cambridge IV supercomputer facility where Berryman conducted his last experiments has been suppressed and classified as highly undesirable knowledge. It's generally believed that Berryman and most of the facility staff died. Subsequently, copies of basilisk B-1 leaked out. This image is famously known as the Parrot for its shape when blurred enough to allow safe viewing. B-1 remains the favorite choice of urban terrorists who use aerosols and stencils to spray basilisk i…

user55340

4:32 PM

Gah! I accidentally formatted this gmail font as comic sans rather than courier new. The pain! What should be fixed with is... gah.

4:42 PM

@JimmyHoffa I've read your answer on Code Review, but... I'm here, and I'm finding it nearly impossible to concentrate. "Lazy loading" right now means coercing someone to bring me another cold one, so I won't have to walk the 20-25 meters to the bar.

Also, your answer brought new attention to the question and now someone's bitching that by saying "brainfuck" I've turned the site into a "hardcore forum" (whatever the hell that is).

user55340

When I worked at SGI, there was a billboard on Highway 101 (the main highway of Silicon Valley) that read "SGI, at the core of business". At that time it was also known that SGI was the server for hosting "Club Love" - one of the early porn sites that had hosted the Pam Anderson sex tape. Internal to the call center we joked they should have redone that billboard to "SGI, at the hard core of business"

@YannisRizos Damn... I should visit Greece sometime

user55340

@JimmyHoffa I've heard stories of lack of productivity of a SillyValley startup located upstairs from a pub. Meetings often involved beer in the "downstairs conference room"

4:59 PM

The Google offices in Cambridge are moving above my (and two of my Googler friends) favorite bar...

user55340

One of my former cow-orkers told me of every bar within a 5 min drive of my new place of work...

@MichaelT An old colleague told me one of his CS professors told all her students software folks have a strong tendency towards habitual drinking, my experience tells me it's a more common distribution than that but being in the company of folks like Phil Katz, who knows

user55340

@JimmyHoffa I haven't had a drop in quite some time... part of it is the convience of having a pub within staggering distance of where I lived. I haven't tried the one near me yet... its also a stress level thing. Previous employment wouldn't have liked me showing up with a hangover.

user55340

SGI tech support had a regular thursday night outing.

user55340

The other bit is I have grey hair in my beard and feel rather uncomfortable in college bar. You want to drink with people who "know your troubles"

5:07 PM

Philips we would have great release parties every month or two, at one large company I worked lunch beers with the team were totally typical

user55340

At Netapp, the releases were named after beer (a very beer culture there...)

user55340

I mean, the software was named Data ONTAP.

user55340

"This is the Heineken release of ONTAP..."

user55340

@GlenH7 btw, the close vote thing... thats people who participate in reviews. There are quite a few active close voters who never review. And unless there's a flag from a <3k user, close vote reviews are at best 80% of close votes because someone has to cast the first in a non-review way.

user41796

@MichaelT Was that for internationalization support?

user55340

5:11 PM

@GlenH7 Dunno... they liked the world wide known brands rather than the more local ones... though I think an early one was Anchor Steam (a SF beer).

user41796

@MichaelT I was taking umbrage with the comment of "there are 5 close voters" in the whole community.

user55340

True. I think if you hunt up some meta questions prior to the last election, there was some data explorer queries that give a different picture. You can find the close votes on all closed questions.

user55340

But only closed questions... failed close votes didn't show up in that query.

user41796

Yes, on that query. I linked to it in my reply to gnat. I think it would be more interesting to see who all has been close voting period. It would also be interesting to see the tallied outcomes of the close review queue votes.

user41796

And I have noticed new questions being fed into the queue automatically with a suggested close reason. But there wasn't a genesis close vote to spark showing up in the review queue.

user55340

5:15 PM

If its in the queue, but no votes on it yet, its from a flag.

Ah didn't know flags would go to queues

user41796

did not know that; thanks for explaining how they got there. had been wondering.

Now I feel better about my pre-3k state I would flag things

user41796

those flags are supposed to show up in the 10k tools for additional verification. Previously 10k tools flags had a special highlight on the menu bar. Haven't seen that in a while though.

@MichaelT you need to go answer something today, pop that 10k hump

Here's a good design question:

1

Q: Event based logging. Is it a good idea and is it ok to pass a handle to a "logged" object?

Hi I have fairly complex program that is doing computations in a quite large loop. I want to log some basic statistic about the run to be able to analyze its performance over time and vs loop number. For now I am outputting a set of numbers, about 5 or 6, each loop. The problem is the number of ...

c++ design-patterns logging software-design

If I had time I'd think up an answer myself

user55340

5:37 PM

@JimmyHoffa I did one this weekend that didn't get too much attention.

user55340

1

A: Create and delay task for 24 hours

I believe your first instinct is the best one - in that it is where you should look for the answer. The early versions of cron was a polling approach (every minute, check to see if anything should be run). A later version made use of a discrete event simulator. The article that lead to this ve...

@MichaelT You were answering the question rather than solving the problem

The guy's problem was he needed to use some cron software and didn't know which one to use, so he asked how to write some cron software

But that's what you're supposed to do

Not your fault he asked the wrong question

user55340

Cron itself wouldn't be the right thing to use. at might be better, if you can shell out, but I'm nor sure how well that would scale.

a lot of those batch software apps are absolutely horrible

user55340

I believe the best approach would be to essentially write a specialized cron for one off tasks meshed in the rest of the system (application).

5:48 PM

I would say so as well, but he would be so far out of his depth

user55340

Understanding how cron works lets you write the software better. And if its done really right, then you can even have multiple readers of the queue to distribute the jobs.

He should be using File Manager Pro and MS access to do complete his software tasks

user55340

I don't think it would be too far out of his depth to write the single thread reader and fork version.

user55340

The multiple system one would get interesting... and you start thinking about message queues again.

Perhaps, but then when's the last time you gave a mildly complex task like that to a junior patted them on their head and sent them on their way, then saw what they wrote a month later? It's...surprising...

user55340

5:50 PM

Store the job in a database, have one job just put messages on the queue when they come due (thats the single threaded version) and then multiple systems each grab a job as it can and process it.

user55340

The single threaded one, after reading the mentioned pdf, I think a Jr. could do it. It would certainly be a good learning experience.

Ah I just read the Q closer, he's not that junior

I've got the best solution ever.

Check it: Messages flow around the system instructing of tasks that need to be completed at time of BLA, you receive this message, write source code out that would sleep every second until that second, then execute the task or send a message to say the task should execute now, compile and run it! :D that's not convoluted at all

user55340

0

Q: Is it sometimes reasonable to cut corners and expect to re-write software in a couple of years?

I work for an organization with one developer (me) and one DBA. When I started, the previous developer had developed applications that had bad architectural practices and it was getting and more time consuming after a couple of years to add features, fix bugs as the story usually goes. I came i...

software software-developer organization software-design

user55340

> I'd expect such applications to be hard to maintain over a long term, and that'd be ok, because we'd just rewrite from scratch and copy and paste any business logic we needed to carry forward - as long as the requirements, documentation, and the data modeling is solid.

user55340

o_O

6:08 PM

do you guys know of anyway to override the member accessor operator in c++? specifically I'm looking to overide the . accessor not the ->

everyone says there isn't a good way, but I need to wrap a non thread safe class with a thread safe accessor, which means overriding its public member variables

@Ampt Override the member accessor operator? Just hit your CPU with a hammer at that point. Kidding, in all reality it's easy, write a Haskell compiler in C++ then you can do it with a simple monad. Just saying...

changing the base non thread safe class isn't really an option

@JimmyHoffa Can't believe I missed that. What an obvious solution lmao

Literally you are trying to make a monad though, which is kind of funny. That's the purpose of the IO monad in haskell, let's you do IO stuff that could be unsafe all wrapped up behind the monad wall so it's unsafety won't leak into your program logic

so I need a monad in c++... hmmm... how long ago did those become mainstream?

bartoszmilewski.com/2011/07/11/monads-in-c

6:12 PM

because I'm working on a legacy system which has all the packages available for debian... in 2007

that's all I get

I'm on boost 1.34.1 lol

Mainstream? Those aren't mainstream right now

It's ok, there's no packages out there that would help you write a monad in C++ anyways so you're not missing anything... ;)

phew, I was worried for a second there. I'll just write my own :D

alright... so no good way to do that. Can I override the actual name then? I know which variables I need to wrap, but it's not like they have parenthesis obviously

Don't ask me, I don't know C++ at all

yapb-soc.blogspot.com/2012/11/monadic-io-in-c.html

@!JimmyHoffa - Can I override the actual name then? I know which variables I need to wrap, but it's not like they have parenthesis obviously

@JimmyHoffa better?

You can't use a decorator, adapter, or facade?

A manager?

6:18 PM

well in essence that's what I'm trying to do, but the class consuming the object accesses the public member variables directly

so I need to step in between there and add a thread safe accessor

Isn't that exactly what decorators are for?

void LinCAN::SendMessage(const CanMessage& message)
{
struct canmsg_t sendmsg={0};

sendmsg.id = message.Identifier;
sendmsg.length = sizeof(message.Data);

memcpy(sendmsg.data, message.Data, sizeof(message.Data));

// send composed message
vca_send_msg_seq(m_FileDescriptor, &sendmsg, 1);
}

its like that

see message.Identifier? I need to override .Identifier

it's not a method

user55340

@GlenH7 ah ha! found it. data.stackexchange.com/programmers/query/90851/top-close-voters

Oh C++ doesn't have property access like field access?

Can you modify that snippet of code to call message.Identifier() instead? Or is that not your code to modify?

this I can't modify :/

neither can I message the CanMessage class

I can only extend it

6:21 PM

C++ really doesn't let you intercept a field access?

not from what I'm seeing lol

In C# you have your public string bla { get { return "arrr!"; } set { _hoyhoy = value; } }

which is a "property" rather than a field

and you would access that property as someclass.bla just as a field

C++ has to have ...something... like that

It's got everything else

oh yes, you can make those methods

but in this instance that's not what happened....

they just accessed the field directly

as a variable as opposed to through a method

methods I can override, but member variables... not so much from what I can see

user55340

@GlenH7 yep, and thats the same one you linked...

@MichaelT man, gnat does not mess around

user55340

6:33 PM

@Ampt he's also on SO's top close vote reviews - stackoverflow.com/review/close/stats

wow lol

user55340

He's also been here for a while...

user55340

Btw, did you see how many outstanding close votes there are in that queue? 60k!

holy CRAP

and only 2k so far today

so theres a months worth of work in the queue

user55340

But yea, there are some who have a rather close oriented view of close votes. You'll also note our downvote rate.

6:36 PM

where do you see that?

user55340

12.5k 9 43 79

java testing project-management agile issue-tracking documentation terminology

user55340

All time - nearly 15k down votes.

wow...

and only 4k up

user55340

I'm only at 1895 down.

that seems like a roundabout way to get what you want

by eliminating everything you don't want

user55340

6:39 PM

It is a way to point users in the right direction. Earlier, I downvoted a user who had a link only answer... asked him to clarify. He clarified, and I changed it to an upvote. It is an important process.

2

this is true as well

user55340

If the answer is wrong, it is important to downvote. Kind of painful at times, but its about the answer and the quality...

2

user55340

Realize though, that this is just one philosphy in the community.

user55340

Personally I prefer close fast, fix (if possible), and reopen. That said, many times the OP doesn't want to fix or really is trying to ask a question that isn't appropriate for P.SE (topic or type of question).

user55340

I'll also go through and find old questions that serve as poor examples and cast close votes on them.

user55340

6:45 PM

There are some other Data Exp.SE queries that I've written that will help find marginal answers - questions that get answers that are < 255 bytes long are often not good questions to begin with.

user20683

7:06 PM

I have 4 downvotes

user20683

technically

user20683

they are from before I switched to my "never downvote" policy.

user20683

@JimmyHoffa

user20683

Enough with Javascript Already: A neat presentation by Nicholas Zakas.

Care to expand on the never downvote policy @WorldEngineer ?

user20683

7:09 PM

@Ampt I understand the intended purpose but I feel like it doesn't explain enough.

user20683

People get really (disproportionately) pissed off about downvotes

So you just never downvote?

user20683

yep

user20683

I even campaigned on it

user20683

and I don't break oaths

7:17 PM

So you would rather see poor questions slide down the front page rather than actively downvoting them? Although do downvotes really do anything other than let the user know that they suck?

it doesn't push them down unless I missed something

user20683

@Ampt at this point, I close or annihilate them

@Ampt It's the hitchikers guide to the galaxy approach to problem solving

user image

that's how I imagine @WorldEngineer when he comes across a post that's not quite great but not bad enough to close

I like that the gif works embedded haha

user20683

@Ampt The Bridge would revolt without them

7:33 PM

@WorldEngineer I don't get it. The guy doesn't like javascript rendered websites because he had lots of clients write huge ones that were bulky and terrible... Ok? But his solution is? Use javascript? Combine it with HTML too? (egh..)

user20683

@JimmyHoffa yeah I'm not sure

user20683

he's the author of the book I'm reading

user20683

so...yeah...

@WorldEngineer What's the book about? Javascript?

user20683

yep

7:37 PM

@JimmyHoffa All of his books are about Javascript

@WorldEngineer What about javascript? You already know it as it is

He writes a lot of Sitepoint tutorials

user20683

@JimmyHoffa I know some basic javascript

user20683

I don't "know" it in that sense

Ah

user20683

7:39 PM

I know it probably about as well as I know C

I agree with him about one thing: People go nuts with the frameworks, you really need a lot less than people use

user55340

@Ampt If you have 3x questions with significant negative rep, the account gets a question ban. Questions and answers with negative reps are candidate for 10k users to delete.

user20683

I think that's what he's really arguing.

I urgently need to find answers for properly testing (including but not limited to unit-testing) for concurrent software. The software is written on top of existing concurrency framework (such as Microsoft ConcRT/PPL/TPL/AAL), but there's still a need to discover latent race conditions and various possible defects. This must have been asked many times over, but keyword search is not a good way to gather all the gems scattered in many questions.

But I disagree that you should be making your javascript HTML aware, I think one of the smartest modern things about javascript is how people are making it html ignorant, it just has some resources it spews around in it's models which happen to create a view, but from the javascript perspective you're just altering your model

7:40 PM

I think people get the idea that if it's there, you have to use it.

user55340

And while he never downvotes, every time he does a close for certain reasons, the reason itself will cast a downvote... and if he's the only one casting it... well, pedantic debates there.

user20683

It hit me last night, most of the JS frameworks out there basically mimic other programming languages

user20683

like underscore.JS looks suspiciously like Ruby.

user20683

object.each() ?

@rwong Bad news, call me crazy, but in my experience, to this day the absolute most effective technique for finding concurrency issues is careful code inspection by talented engineers (who understand concurrency well enough to know what they're looking for)

That said you can still create tests and use other techniques, that's just the most effective

7:42 PM

@WorldEngineer And then there's this: pyjs.org

for that particular domain

user20683

@Dynamic Brython is the better known one

user20683

brython.info

user55340

@rwong Real world concurrency doesn't always show up in tests. Significant logging of the state when you have trouble in production is also needed. Don't think you can automate tests to some comfort level... you'll never get it with concurrency.

@WorldEngineer There's an entire list of languages that compile to JS:

github.com/jashkenas/coffee-script/wiki/…

Most of them are harder to use/learn than Javascript itself.

user20683

7:47 PM

people and their petrification of learning new languages

user20683

>.>

Actors model is the one I love and am confident of a low-defect implementation. However, in my application actors model only uses 50%-60% of the available concurrency (throw in 4 physical cores, it will only use around 2-3 cores) because of the imbalanced nature of stages. Switching to a task-decomposition model (which I've done and now in need of test) adds significantly more lines-of-code (~5 times more than actors model!) which I lost any confidence whatsoever.

@WorldEngineer A lot of frameworks stuff like .each are to work around danger holes in javascript itself, it's easy to make minor mistakes in javascript in certain places you don't realize that have ramifications you don't expect, higher abstractions allow that stuff to be safety-scissored

user20683

yeah

That said, a lot of the javascript frameworks seem to me more about giving people who aren't altogether knowledgeable engineers the ability to just magic up a website without any idea what they're doing

7:48 PM

@WorldEngineer There's extremists with everything

user20683

I got nailed by "for...in" last night

user20683

I was like "This should work like Python"

user20683

it does not work like Python.

@rwong I agree, actors model is great and significantly reduces your defect rate

(As a side note, I'm really surprised that the concurrency::transformer implementation in Microsoft ConcRT doesn't take care of load-balancing (creating multiple agents), leading to under-utilization of concurrency.)

7:52 PM

Oh man you're working in C++, concurrency is a far scarier world there I would imagine, deterministic destruction is the only + point for concurrency being easier there, but the meddling with memory everywhere and pointers as a standard makes it wayy scarier

Is there any website with news specifically about software libre (free software) except fsf.org?

I learned that lack of Garbage Collection makes concurrency 100% more defect-ful.

@Derfder Look:

public static void main() { Console.WriteLine("Hello world!"); }

You're bit of free software for the day!

as in, half of race conditions happen in destructor :)

@JimmyHoffa except your Hello world example? Do you know any website that write articles about free software?

steve_gallagher

7:56 PM

Trying to make a list of canonical application types, i.e : To-do list, Shopping Cart, Twitter clone, Travel Reservation system, Blog, etc. Does anyone else have any ideas?

@Derfder for industry news, see infoq.com
for anything else .. please be more specific.

steve_gallagher

Simple straight-forward apps that you can practice a new language with or use as a demo for competency

@steve_gallagher how about this: CRUD, server-client system, content distribution system, realtime messaging, broadcast system, publish-subscribe system, transaction system ...

steve_gallagher

@rwong, thats beautiful thanks

@steve_gallagher What about the application type that creates canonical lists? Or the application type that creates applications that create canonical lists? Or the application type that creates applications that create applications that create canonical lists?

user20683

7:58 PM

@JimmyHoffa Stop huffing the recursion

thorsten müller

on a lower level: a sudoku solver, conway's game of life

« first day (1046 days earlier) ← previous day next day → last day (3950 days later) »

Transcript for

Jul '1315

The Whiteboard

General Discussion for softwareengineering.stackexchange.com P...

coffee-day horror-stories monads my-code-is-compiling scotch the-other-monitor

join this room
about this room

00:00

06:00

12:00

18:00

all times are UTC

site design / logo © 2024 Stack Exchange Inc; legal

mobile