« first day (1890 days earlier)      last day (3109 days later) » 

user41796
5:00 PM
And C# is MS's answer to Java, so ...
 
@Telastyn yeah, that's what my thoughts are. It's not bad, but there are better options already. The type system lacks an amount of flexibility I would like, the language is more verbose than is necessary; one of the features I would love to see would be Go's style implicit interfaces. Constraint-inferred generics would be nice as well
 
As a beginner, one of the big issues I had (and still have) was picking a language I could spend a lot of time with learning the trade
 
Tangent started out with Go implicit interfaces, but that leads to a lot of tradeoffs elsewhere in the language. I can live without them now I think.
 
I tried a few, and I kinda enjoyed C# and Ruby a little more than the others
 
yeah, I don't want true functional
but just cutting some of the C# fat would be nice.
 
5:01 PM
@Telastyn lots of these things lead to tons of tradeoffs; the constraint-inferred generics are basically tossered just by the fact that the language uses subtypal inheritance.
 
Do people see Ruby sticking around for quite some time?
 
true
 
@daOnlyBG probably, though it's a smaller portion of current new development than many other languages. Choosing your language is a difficult thing because you have to know what you're aiming for: $$$ ? Longevity? Enjoyment? Opportunity?
 
It's a good question. The naive answer would be to say "all of them," but I understand there are tradeoffs
 
@daOnlyBG no language that has crossed a certain threshold of popularity (= has lots of legacy code) is going to die soon. There's till tons of Fortran and Cobol stuff around. In the Tiobe-top-X index, only Objective-C is going away soon, because Apple is pushing a newer language
 
5:04 PM
Ah yes, "Swift"
....I didn't like objective-c...
My only fear is spending a lot of time in one language, just to finally have prospective employers think "Well, we're not really going to use this language any more, etc"
but maybe that's an irrational fear. Idk
 
@amon when people talk about dying though there's a threshold of new development happening in them that they're really referring to. I think most people know legacy software will always be there and will always need maintenance, but there's very little new development happening in COBOL despite the massive amounts of maintenance and still active systems running in it today
 
all languages will have that happen during the duration of your career.
it will happen more than once even.
 
user41796
And now I'm out of close votes for the day
 
Ah, COBOL. Mainframe technology, right?
 
> 17 votes remaining
 
5:08 PM
C# will hit that threshold by 20 years from now I believe. 10 years? Iduno; plausibly. I hope so. There is a huge amount of people moving to Node.JS these days- more and more .NET is being relegated to the desktop applications space as server side stuff is done in JavaScript
 
do you think desktop apps are going to continue to migrate to cloud based apps?
 
Corporate software aside, yes
 
I'm curious even about corporate software though
 
@enderland absolutely; it's the only approach to cross platforming that's showing promise and as mobile hardware improves and browsers alongside them... Browsers are effectively the modern OS
 
node.js is a tire fire.
 
5:10 PM
because for us, we have huge tech debt associated with the pain in upgrading software versions/etc - this could largely be mitigated if you were running the applications exclusively on a cloud based setup
 
(I'm not saying it wouldn't happen for corporate software, I'm just unsure about that area, and more sure about personal software, etc)
 
@Telastyn only because the skill level of those living in it... it's perfectly good for people coming to it with a solid background in software design and analysis (a minority of folks doing Node.JS)
 
naw.
 
you don't think people can write good clean JavaScript?
 
let's use a 1 week programming language without concurrency support or type safety for running services? Yeah, that's going to end well.
 
5:12 PM
So I'm going through a bunch of state government websites. Guess what percentage of them use tables for layout?
 
98%?
 
user114359
102%
 
yep, government websites...lol
 
@Telastyn you have to look at Node.JS as a scripting language used to script a highly concurrent asynchronous runtime environment. The runtime environment is written in C++ and has shown to be very efficient, the JavaScript is just the scripting language with which you operate that runtime environment. People doing data crunching with the scripting language rather than through the runtime environment's call-outs are failing multiple sensible-design tests...
 
user41796
@RobertHarvey Guilty secret - sometimes I have resorted to using tables when I've needed a neatly aligned grid layout.
 
user41796
5:15 PM
We pull in a lot of key-value-pair type data, and a grid is a great? easy way to lay that out cleanly.
 
@GlenH7 cut cut cut! Seriously though, learn a grid system. With a good grid system HTML layouts become so damned much simpler.
 
@GlenH7 CSS display: table; /* all of the layout, none of the semantic baggage */
 
@GlenH7 if you have tabular data to represent, ok, use a table. That's what it's for. If you have a bunch of data you want to present as a spreadsheet? Great.
 
user41796
Truth be told, a lot of our user base prefers tabular presentation of information and data entry. Just the nature of the beast.
 
Would you fine folks entertain a computer hardware question? I'm primarily a Mac OSX user, but thinking of getting a mini computer for desk at home, and had some questions about certain processor technologies
 
user41796
5:31 PM
Just get a mac mini and be done. :-)
 
user55340
Or air.
 
@GlenH7 booooooo. your recommendations are bad and you should feel bad
 
Heh, I was thinking of getting a small machine that I could install both Windows and Ubuntu
I'm just not a big fan of Xamarin for Mac OSX
So one candidate was the Intel NUC
 
@daOnlyBG Surface Book looks sweeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeet
 
user41796
2 days ago, by Jimmy Hoffa
@GlenH7 SCOTCH YOU!
 
5:34 PM
but then I heard about the ARM processor
However, ARM processors don't exactly support windows technology at the moment- am I understanding that correctly...?
 
user41796
@daOnlyBG Surface RT / Win RT is dead.
 
Well, that matter is settled, then
Beside the price, what other objections would one have toward the Intel NUC?
 
user41796
You could argue WinRT was DOA, but that's just splitting hairs. It's dead now.
 
user41796
Classic battle between Intel and AMD is raw performance vs. number of cores
 
@daOnlyBG how is the price an objection? I've known multiple people who picked up a NUC and all of them loved it. I always build my own shit for cheaper though personally. Amazon and Ebay for parts; takes an hour or two one night to put it together and saves me hundreds
 
user41796
5:38 PM
Intel generally runs smaller numbers of cores, but faster and with excellent floating point support.
 
user41796
AMD throws more cores at things instead.
 
user41796
Intel is great for calculation type operations, whereas AMD is great for virtualization
 
@GlenH7 intel sadly still wins on compatibility... it shouldn't even be a thing anymore, but it still is.
 
user41796
that's a good point. It's always been the defacto processor on the Winders side
 
it's annoyingthat it's faster to call a hotel directly and make reservations than go through our online booking tool
 
user41796
5:40 PM
In full disclosure, I'm running an AMD chip in my HTPC
 
Every time I think I can buy AMD or ATI and not have to worry about hardware or driver compatibility because it's not '95, I come up wrong. With prices on intel and nvidia hardware that's solid as can be these days so good, there's no longer even really a monetary reason to opt other than the standard bits
 
user41796
@whatsisname painfully true
 
user41796
1 hour ago, by GlenH7
preemptive declaration that animated gifs may be deleted immediately after posting
 
animated gifs are the greatest thing ever created
 
5:43 PM
^-- $55 - high performance for practicaly no $$
 
user41796
@whatsisname I think that one is ready to go into it's history though. It's hilarious, but I'm already tired of seeing it.
 
the cheapest CPUs these days are more than capable of doing bloody well anything and everything you ask of them
 
user55340
@GlenH7 check your close review count.
 
user41796
@MichaelT daaaaaaaaaaaaaaaaamn
 
user41796
5:45 PM
user image
3
 
@JimmyHoffa I should clarify: I can def afford the NUC, it's not really that expensive IMO. I just wanted to set price aside in case anyone else thought it was too expensive. I could have typed "Assume price isn't an objection"
 
Day 4: still no news from the front
 
(which probably would have been more appropriate)
 
user41796
@enderland Did they give you any info on when to expect things?
 
user41796
Give it till 2 or 3 PM today, then drop a note
 
5:48 PM
@daOnlyBG you want performance on a budget? AMD/ATI. You want the best money can buy? Intel/Nvidia
 
lol
 
My primary objective for the hardware is to do math&stats calculation, primarily monte carlo simulation. So I guess I'd be leaning toward the Intel processor. I assume it's worthwhile to chase the multi-threading capabilities?
 
user41796
@daOnlyBG with monte carlo? Absolutely
 
@daOnlyBG the NUC is intentionally sold as a cheap option
 
@daOnlyBG is price a consideration? Or is this a professional thing?
 
5:49 PM
are you saying that professionals don't care about price? :)
 
Well, it'll be coming out of my own pocket- I'm working on a project that will prob take 1-2 years to do
 
Honestly you may want to look at AMD just because you will get more bang for your buck
 
but in the end, I hope it will help my professional career
 
user41796
Do you have an MSDN subscription? As that gets you free $$ with Azure
 
nah
(not yet)
 
5:50 PM
> I hope it will end my professional career
reading comprehension ftl :(
 
user41796
Also worth looking at AWS as it can simplify some things for you
 
user41796
@enderland Writing the next Flappy Birds could end my professional career
 
I thought that's what it said at first. lol
 
I really need to start blogging. I could become famous. Because if I become famous at my day job, that's...not a good thing.
 
are you being serious @ThomasOwens?
 
5:52 PM
@enderland Yes.
 
@ThomasOwens Like... live in russia famous?
I kid :)
 
@Ampt bah. AMD/ATI isn't really particularly cheaper these days... have a look here:
 
@GlenH7 what would AWS simplify?
 
@JimmyHoffa he's only looking at CPUs
 
@Ampt Well, that's how I could become famous at my day job. And that's not the kind of famous I want to be. In fact, I want to be the total opposite of that famous. I very much like my day job (85% of the time, at least).
 
@Ampt I'm just pointing out. I already linked a $55 intel CPU that plays every game I throw at it..
 
user41796
@daOnlyBG worrying about handling multiple runs or buying a big enough processor to make runs complete in a reasonable period of time
 
@JimmyHoffa he's not gaming either - he's running simulations
I am 100% an intel/Nvidia guy, but I understand that AMD/ATI have their niche markets
 
Oh yeah. I never got into gaming. All my friends did, though. I picked up the saxophone instead.
 
@Ampt I still don't buy the "AMD is cheaper than intel" bit anymore with how cheap intel stuff is these days
 
5:54 PM
(Nothing against it, it just never materialized with me, etc)
 
and cranking out crazy numbers for half the price is one of them
 
@Ampt that argument held until maybe 3-4 years ago. Intel's value options just make AMD look silly these days though..
 
the benchmarks don't lie
that last one is particualrly damning
 
@Ampt that isn't a rational choice though because it's not weighted to cheap - you can get cheaper options from both Intel and AMD
 
@JimmyHoffa I'm weighting towards price/performance while trying to maximize performance
 
5:57 PM
Do you sacrifice anything by going for cheaper options from Intel?
(not quality, but performance, etc)
 
@daOnlyBG just performance. it's pretty linear up til you get to the high end stuff
then it gets exponential, quickly
 
how "high end" are we talking about? the i7 core processors?
 
anything with more than 4 cores and it starts go get expensive, very quickly
 
note the price of all choices around the performance of what I just linked
 
6:00 PM
Hmm.. pretty useful webpage
 
@daOnlyBG I don't think so honestly. I disagree with most on this as I've run cheapo CPUs for years and never had issue because desktop use software doesn't do shit on the CPU really. Not enough to stress a modern CPU. Get fast ram, get fast SSD, and if you do anything graphically intense get a GPU to offload it; but CPU's in desktop uses are so not stressed these days
 
@JimmyHoffa yeah, I probably wouldn't go for something that strong anyway. I figure if I'm going to need an i7, I should probably be doing such computations on a bigger computer at work/university
there is an i5 that offers multithreading technology, though,
and I'm wondering if it'd be worth the time/money buying it
So for someone running monte carlo simulations- what would I have to do to pretty much tire the processor?
 
tire the processor?
 
push it to the limit
 
6:03 PM
multithread the app n times, where n is the number of virtual CPUS
 
justify the need for a stronger one, etc
 
you'd have to ask @GlenH7 that one
 
user41796
@Ampt That would do it actually
 
Whatever server My Documents is on is incredibly slow. Our SharePoint servers are also not working. I have access to almost 0 of what I need. But I do have access to some things.
 
user41796
monte carlo is an example of a stupidly parallel operation. Stupidly parallel meaning it's easy to convert the calculation to something that can be run in parallel
 
6:05 PM
uh huh
 
user41796
10,000 runs of essentially the same calculation with slightly different inputs begs for being run in parallel
 
seriously though, for 170 bucks, you get the performance of a 300 dollar i7
 
user55340
In parallel computing, an embarrassingly parallel workload, or embarrassingly parallel problem, is one for which little or no effort is required to separate the problem into a number of parallel tasks. This is often the case where there exists no dependency (or communication) between those parallel tasks. Embarrassingly parallel problems (also called "perfectly parallel" or "pleasingly parallel") tend to require little or no communication of results between tasks, and are thus different from distributed computing problems that require communication between tasks, especially communication of...
 
@MichaelT I bet those workloads show up to high school in their underwear too!
 
why would someone still argue for an Intel, then?
nvm- I think I might be nitpicking a few things unnecessarily. Thanks for the advice!
 
6:20 PM
@daOnlyBG at very low prices (sub 60 bucks,) the trend kind of stops, and intel/AMD are equal for the most part
but for the 200->100 dollar range, AMD wins, which is on purpose. That's really AMDs sweet spot
Does your monte carlo simulation use heavy floating point?
 
0
Q: How to determine if there's something wrong with my code or a bug in the library

ahotaI'm working with Meteor, which is a relatively young Javascript framework. As such, it seems like I often run into problems which haven't been asked on Stackoverflow already. In one such case, I was agonizing over something that wasn't working, and it turns out it was a bug with Meteor. When I ...

that seems like a good question (probably a dup somewhere though)
 
user55340
One more vote please: and then 20ks
 
user55340
-6
Q: Text to speech sound record and database

Noor Alarefmy graduation project about text to speech (TTS) , i'm looking for programs to record the voice and then convert it into Database , is there any program can help me or not ??? if there is no program like that , how can I create a database for my record voice?

 
user55340
Thank you. Now for @GlenH7 and @RobertHarvey
 
@Ampt- to answer your question, yeah, they can be heavy floating points
 
6:31 PM
@daOnlyBG hmmm... you may want to investigate what benchmarks take that into account, and look for results based on those
 
I most definitely will
 
y no expand?
 
fixed. :)
 
:D
 
needs to end in .jpg / etc
 
user41796
@MichaelT BOOM!
 
user55340
@GlenH7 lot of therapy opportunities out there if you want them.
 
user55340
I found a treasure trove stinking cesspit last night.
 
user41796
I'd believe that
 
Thanks. It would be great if you can give a real-life example. — Agnib 9 mins ago
[headdesk]
 
I'll get right on that.
 
user55340
@enderland tweaked your answer. Feel free to revert.
 
psr
@JimmyHoffa Set up a high resolution video camera so you can better make out whether or not the intruder has a camera.
 
@MichaelT nah that' fine
 
user41796
@enderland <polite cough/>
 
@enderland "In public interest (for lots of money)"
 
It is one of the more well written "is software engineering engineering?" articles I've read, though somewhat intentionally provacative
> But fifty years’ worth of attempts to turn software development into a legitimate engineering practice have failed.
 
@enderland PE licensing group says yes, so I'm inclined to agree.
The title and subtitle alone turned me off to the point that I'm not going to be able to fairly give the article a chance.
 
same
 
7:25 PM
I do agree with the statement that the title of engineer is being "cheapened by the tech industry". A lot of companies are giving people the role title of "software engineer" or "dev ops engineer" or "systems engineer" or whatever. Software development should learn a lot from engineering principles, but not all software development is engineering and companies that don't require the rigor of engineering shouldn't call their people engineers.
 
@ThomasOwens I think this is the point the article is making, and I think it does make it pretty well
 
that was on slashdot too
 
user41796
devops engineer is an oxymoron
 
@enderland The article seems to say that the agile methods aren't engineering. I guess I agree with the article at a very high level. But not with some of the points.
Some of the specifics are...not right.
 
the majority of software development isn't 'engineering' and it shouldn't be
 
user55340
 
It also equates computer science education with software engineering education, which I think is a huge mistake. They are distinct fields, even though a relatively small number of school have software engineering programs.
 
@ThomasOwens More and more CS programs are splitting into SE and CS
 
@ThomasOwens I have to agree with your sentiments. If you see the title my company gave me when I joined, you'd laugh out loud- and I'm not even selling myself short
 
In the traditional engineering disciplines, what do you call people who aren't engineers but do work?
 
technicians
 
user55340
7:29 PM
SE means "coder" in that context. Not engineer.
 
if by "aren't engineers' you mean "arent engineers in terms of work" and not "aren't engineers because no PE"
 
@whatsisname See. That's a good word for it. Software Technician. Or something along those lines. It sounds good.
 
would it be bad to ask what "engineer" actually means in the context of this conversation? it's becoming very obvious that I simply don't know
 
no way
 
even "analyst" would be more appropriate than "engineer" the vast majority of the time the latter is used
 
7:30 PM
@Ixrec What's confusing about the term?
> An engineer is a professional practitioner of engineering, concerned with applying scientific knowledge, mathematics, and ingenuity to develop solutions for technical, societal and commercial problems.
 
@Ixrec: there are 4 kinds of definitions for engineer:
 
@Ampt garbage - it's age makes it incompatible with faster ram and mobo technologies that make the $55 CPU I referenced a better option
 
that definition comes close to being circular
 
1) people who do technical work that uses science to get your stuff done
2) people that follow engineering methodology in their work
3) people that are engineers as per a licensing/certification board
4) people that work on computers
 
lol
 
7:31 PM
5) people that operate train locomotives
the conflict between those definitions leads to stuff like the article being discussed
 
I think most people have a reverence for what engineers do that makes them assume that everything that is "engineered" goes through some magical process making it perfect
 
I am 1 and 2. I would be 3 today, but there was no FE exam I could pass and the PE exam wasn't even out yet when I Was in school.
 
could you elaborate on #1 and #2? those seem like the relevant ones
 
1: "magic"
2: SWAG
 
while I wouldn't use those phrases to describe what I do, it's also not immediately obvious that what I do is so lacking in "science" or "methodology" that calling it software engineering would cheapen the term
 
7:33 PM
@Ixrec: #2 is people that often care more about the process of getting there than what results
 
@Ixrec 1 implies a grounding in mathematics and science. For mechanical engineers, physics, for example. For software engineers, discrete mathematics and computer science (like, true mathematical computer science).
 
people that think a lot about ISO standards and IEEE swebok nonsense
 
@JimmyHoffa It uses DDR3, has the AM3+ socket (a still used mobo socket) and has a faster benchmark than the processor you linked.
 
#1 is the more casual concept, someone who does work that answers only to the laws of physics
 
@ThomasOwens would simple big-O calculations qualify or do you have more advanced stuff in mind?
 
7:34 PM
@ThomasOwens what percentage of ME's do you think use math/physics on a more regular basis than software engineers?
 
@Ixrec That's part of it.
 
@ThomasOwens you can get your PE now if you want
 
Optimization would be another part, I'd imagine
 
@enderland What do you mean? An ME building something must conform to the rules of physics. They must understand how their thing will behave in the real world.
@Ampt No, it's not feasible. I looked.
 
7:35 PM
a lot of MechEs eyeball stuff or just look stuff up in tables
with minimal computations done
 
@ThomasOwens I mean, what percentage of them are using anything other than intuition and experience in doing any design work?
 
@JimmyHoffa did you even scroll down to the benchmarks?
 
@whatsisname That could be, but the underpinnings of what they do are in physics.
 
in the software I work on optimization mostly consists of measuring how long things take, and knowing the system well enough to say "this thing is taking way longer than it should"
not a lot of math involved
 
95% of MechEs are doing HVAC and making mounting brackets
 
7:35 PM
the 6300 trounces it in every test
 
the only difference between most ME's and software dev is that you can physically create products and test them easier than software
 
@enderland I... uh.. what?
 
and laws of physics are just as real for software, they just happen to be things like the speed of electrons rather than newtonian mechanics
 
Don't forget - I'm sitting here in aerospace. We also manufacture. So our mechanical engineers build structures that hold optics and withstand stresses associated with flying.
 
@enderland: I think it's the exact opposite of that
 
7:36 PM
@Ixrec That's more of a computer engineering thing.
 
@whatsisname I don't, most consumer products that you buy work and are not physically "buggy" the way software is
 
@JimmyHoffa even more damning - the 6300 lets you use BETTER ram than the brand new intel one
 
@enderland: we would have a much different way of doing things if someone could snap their fingers and a bridge would be instantly built, for free where its specified
we would build bridges like we make software if we could do that
 
@whatsisname just to check, was this meant seriously or sarcastically?
 
@Ixrec: disagree, the laws of physics for software are for the most part non-existent for the vast majority of software develors
@Ixrec: seriously
part of the reason so much software is made so 'sloppily' compared to physical goods is because software is trivial to manufacture
 
7:39 PM
I'll take that as "I've run out of empirical evidence, so here's a flashy gif"
:P
 
I guess we mean different things by that, to me things like "all machines have finite memory and CPU speed" count as physical laws
 
user55340
It's because the defects are hard to pinpoint.
 
@MichaelT your pins just too big is all
 
yes, there are physical laws that govern the computer, but very few of us go up against those
 
@Ixrec To me, that's not a physical law. That's a constraint.
 
7:40 PM
when I make a lot of the client software we use, I just pretend the computer has unlimited harddrive space and memory and never think about it
and I have never, ever run into a problem by thinking that way
there is no builder than can just assume he has unlimited concrete
 
I honestly don't care about the physical laws that govern memory capacity or CPU speed. I care about the hardware that's available to me.
 
@Ampt I concede, apparently AMD does have reasonable choices at that price point. Though I'm still sticking with my "Totally unnecessary for typical desktop activities" point - but he's doing monte carlo simulations...so...yeah. Just about nobody is doing highly paralell CPU work on a desktop outside of a dev
 
The computer engineers can work to make faster processors, larger memory chips and so on.
 
inside the execution land of the software, it takes place in a magical world thata is defined by our concepts of what we want it to be more than anything
 
@JimmyHoffa I maintain that for mid-level desktops, the price to performance of AMD/ATI offerings just can't be beat
 
7:41 PM
and thats a brave new world for 'engineering', there are very few physical constraints we have to work with
 
ok, I think I get the idea now
 
more often the challenge is "is this form sensical to the user" rather than "will this collapse"
 
My constraints are more theoretical. I have data to sort. I have a bunch of algorithms developed by computer scientists. Given a particular data set, there's an absolute best possible approach to sort it.
 
MSWord does not see very many constraints imposed by physics
 
Of course, the real-world performance of that is going to depend on the hardware that the software is run on. But I care that I choose the right algorithm to sort my data.
 
7:43 PM
@whatsisname bullshit, I once saw an MSWord box run over by a car- those disks never worked the same again. Physics: 1 MSWord: 0
2
 
lol
 
would something like developing the concurrent garbage collector for Go qualify as "software engineering"? I'm sure that involves a heck of a lot more rigor, methodology and statistical analysis than what I do
 
@JimmyHoffa was that MSWord box dropped deliberately onto the street, or calmly placed down? If the former, more like Physics 2, MSWord 0
 
so, it doesn't make sense to consider developing MSword to be an engineering project
 
@Ixrec I'd call that computer science.
 
7:44 PM
ok
 
@Ixrec: there are software things that do bump up against those constraints, where laws of information do apply, but those are small
 
@whatsisname Probably not. Especially today. The article makes a point where, at one time, getting patches to customers was hard. You needed more rigor to make sure that you had few defects.
 
while some people get to write compilers, most people are doing CRUD apps for dentist offices
 
user55340
So beyond the acknowledgement that reading in a fixed number of columns doesn't work, what have you tried? I don't know if an array is the best answer, or a vector, or if some other more complex data structure, or if its just a simple finite state automata that chugs along until the end of the line. You need to describe the problem that you are having better along with what you have tried and discounted. Until you describe these things in the post, we can't say what the proper solution to your problem is. — MichaelT 1 min ago
 
@ThomasOwens: those kinds of things are issues that aren't engineering, in a no-true-scotsman sort of way
 
7:46 PM
@whatsisname Cost is definitely a concern of engineers.
 
yes, however keeping tabs on the factory workers and keeping them from screwing things up and keeping manufacturing defects down is a different area
the software engineering <-> physical engineering analogy is already breaking down
 
@whatsisname Maybe in manufacturing.
 
@Ixrec does it? IME the most complex code I've worked with has always had a focus on being easily reasoned out because once you get down to it - most have little faith in formal rigor and methodology compared to the faith they have in clear and reasonable logical exposition
 
software doesn't really have a distinction between defects in design and defects in manufacture
 
Software engineering is unique, at least how it's implemented. In other engineering disciplines where I work, you have "design engineering" and "manufacturing engineering". Software engineering both does the design, but the "manufacturing" steps - writing code, build scripts, and producing executables.
 
7:49 PM
yes, software has nearly free cost and essentially perfect manufacturing
 
@whatsisname I'd agree with this. Especially since code is design. The only "defect in manufacture" would be something that prevents identical copies from being made.
 
which is much different than other fields
where challenges of manufacturing your design dominate
 
@JimmyHoffa well, I've never worked on the implementation of a language used by hundreds or thousands of other developers, but I assume anyone who is tries to be much more careful with their changes than I am with mine
of course that's an extreme minority
 
if you can remove a screw from an engine in a GM car, it's worth paying two engineers for an entire year to do just that
 
@whatsisname There are some interesting design challenges, though.
 
7:51 PM
that doesn't happen in software for the most part
 
@Ixrec true, I'm sure, but I suspect the process is less about utilized formalisms and more about the same process used for mathematical systems: Design, implement, document, review, review, review
 
there definitely are challenges, they are just much different
 
The screw is weight. The other engineering disciplines are concerned with size, weight, and power in some environments.
Software doesn't weigh anything.
 
@JimmyHoffa I guess if you're talking about standardized formalisms rather than procedures that the team themselves decided on, yeah I'd agree
 
no, the screw is cost of the screw and the difficulty of installing it at the factory
if the screw weighed nothing my claim would still be true
 
7:53 PM
That is true. Well, the cost of the screw is always true. The difficulty varies.
Depends on where it is.
Maybe paying the engineers to find a way to move the screw would be worth it.
 
to be fair, it's common to spend a significant chunk of development time to remove steps from an algorithm so that the computer can do less work but get the same result
which is roughly analogous to the screw thing but I see your point anyway
 
That's why DFX is important, where X is "testability" or "manufacturability" or "assembly" or "cost"...
 
no, that's not analogous to the engine screw
 
@whatsisname Why not?
 
@ThomasOwens what's DF*?
 
7:54 PM
@Ixrec "Design for X".
 
@Ixrec no you don't, he wasn't making a point, and neither are you or any of you! It's Friday on The Whiteboard and you're all speaking about as coherently as a category theorist
 
Where X is some factor that is important to you. Reliability, Manufacturability, Test, are big ones. Cost is important, too.
 
because computers don't require wages and benefits and bathrooms and floor space like humans do, humans which have to place the screws in the engine and tighten them to tight tolerances
 
@JimmyHoffa the entire premise of this conversation is debating what "engineer" means, what more did you expect? =)
 
and that have to do it for every engine that rolls off the line
 
user41796
7:55 PM
Acquaintance of mine used to work for Ford. He said one of the things he hated was having his designs going through accounting for review. They'd be ruthless in attempting to strip out as many parts as possible.
 
you think 500 bucks for a copy of visual studio is expensive?
try 500,000 dollars for an injection mold for a vacuum cleaner housing
 
@whatsisname Computers do require floor space. We put equipment in a mobile truck. Being able to do the same processing on a smaller server that requires less power, puts out less heat, or takes up less room would be worth it.
 
user41796
Data center costs are definitely evaluated by the larger companies
 
Although, if things like that are a concern to you, then you're probably doing software engineering.
 
computers also require admins and various support tooling to manage deployments and recover from hardware faults and all sorts of other nonsense
 
7:57 PM
@Ixrec and you're all talking about whether a theoretical screw exists or not in an ideal world where the planet is a perfect circle and the firmament above is a flat plane and whether those theoretical non-non-existent screws have threads that may or may not cost bathroom breaks.
 
yes, computers require floor space, but a computers physical space, power, hvac implications don't even come close to matching that of supporting an equal number of people
I can order 50 computers and set them up in the workshop next week, I could not bring in 50 new employees in our space, even if we had the money
 
and that's why computers get all the good jobs nowadays
 
@whatsisname But they do in some cases. How about another example. We're putting a system on an aircraft. Being able to do the same software processing with 2 fewer boards means less power consumption, less weight on the aircraft, and therefore better aircraft performance.
 

« first day (1890 days earlier)      last day (3109 days later) »