« first day (1518 days earlier)      last day (3459 days later) » 

enderland
enderland
12:23 AM
@MichaelT so far for what?
 
user55340
@enderland 50 to badge.
 
enderland
enderland
Ooooh, Repcap, not recap.
 
user55340
> Earned 200 daily reputation 50 times
 
enderland
enderland
I wonder how close to that I am on workplace, probably not that close
 
user55340
Go to /reputation.
 
enderland
enderland
12:25 AM
29 it looks like
 
user55340
 
enderland
enderland
18 with 200, 9 with 215, and 2 with 230 (and none more)
@MichaelT yeah that's what I was using
I guess I'm halfway there ha
 
user55340
12:44 AM
Huh... The Box of Delights... one of the actors is Patrick Troughton (aka 2nd doctor)
 
Jimmy Hoffa
Jimmy Hoffa
is anybody ready to throw in the towel with the latest Doctor? My wife is so done with this jackass
I'm hopeful, but it may be misplaced faith...
mostly just hopeful they'll fix this busted ass character and writing experiment they're on about
fellow at work says they've done mean doctors in the past like this, but he also said the show went downhill and died not long after such
(surprise surprise)
 
user55340
I'm still of the opinion that he's an interesting change to the character.
 
Stack Exchange
Stack Exchange
Programming in the 21st Century: Extreme Formatting
posted on October 29, 2014 by James Hague

Not long after I wrote Solving the Wrong Problem, it occurred to me that this site is small because of what I decided to leave out, and that I never tried to optimize what remained. To that end I used a PNG reducer on the two images (one accompanies Accidental Innovation, the other is in The End is Near for Vertical Tab). And I ran the CSS through a web-based minifier. A Cascading Style Sheet…

 
Jimmy Hoffa
Jimmy Hoffa
@MichaelT I won't say it's not interesting, I just don't think it's entertaining, which is what I always watched the show for: Entertainment. They're trying to be meaningful and more than just a fun show to watch, but... a fun show to watch is been the point from the 2005 series inception...
 
user55340
There is a "you have to make the hard choices" and it won't always be "the doctor will fix things for you"
 
Jimmy Hoffa
Jimmy Hoffa
12:53 AM
@MichaelT which is not what the vast majority of the current audience is watching the show for though; my biggest fear is the show's simply going to tank because of this. It's not what the audience wants.
 
user55340
I'm quite curious where this story arc is going.
 
user15026
@JimmyHoffa I am actually more into Doctor Who with the last few episodes than I have been for a while now
 
user55340
They're noting that the second episode was down to 1.59M from the season opening (a 23.4% drop). However, that should be perspective of Season 7 premier was only 1.6M.
 
user55340
It can be awkward to get numbers, and I might be comparing the wrong ones... ibtimes.com/…
 
user55340
There also seems to be some UK / US culture issues in there... that he's doing even better in the UK than the US.
 
user55340
12:59 AM
Rotten Tomatoes suggests some good numbers. rottentomatoes.com/tv/doctor-who/s08
 
user55340
Season 8 also seems to be twisting into a very interesting bit with the arc... (spoilers at doctorwhotv.co.uk/… )
 
user55340
btw, @AshleyNunn today is (was) apparently National Cat day in the US. nationalcatday.com
 
user55340
National Cat Day
National Cat Day was founded by animal welfare advocate Colleen Paige. It is a celebration that takes place on October 29th every year in the United States. The National Cat Day website states that the holiday was first celebrated in 2005 "to help galvanize the public to recognize the number of cats that need to be rescued each year and also to encourage cat lovers to celebrate the cat(s) in their life for the unconditional love and companionship they bestow upon us." The day is supported by the American Society for the Prevention of Cruelty to Animals, a nonprofit organization which also works...
 
user15026
1:16 AM
Huh, neat
 
Jimmy Hoffa
Jimmy Hoffa
2:35 AM
@amon I don't necessarily agree with your statement that such observations wouldn't help him determine the color of a square. Say for instance all squares are green, then it would be quite apt to say that squares are not yellow, they are green, and we could then put his mind at ease by letting him know, that in fact, the square he's looking at is green. — Jimmy Hoffa 6 secs ago
@amon ...I might be trolling a bit much sometimes... I can't help myself
 
 
8 hours later…
maple_shaft
maple_shaft
10:53 AM
@JimmyHoffa Maybe I am not explaining it well enough or giving good context. The bank has a number of mainframe systems that are SOR (System of Record) for various types of data, Eg. Name/Address, account validation, etc...
These Mainframe systems provide an interface to the world through a CICS gateway (WebSphere MQ).
In a perfect world a lookup service like Name/Address for all bank customers would be a synchronous one (Eg. Request/Reply Architectural pattern, SOAP, REST, etc...)
The only means I have to talk to this world is through this MQ interface which is asynchronous. The model here is how one would typically do request-reply with MQ. Send messages to the request queue with a correllation ID, then have another process/thread listening on the reply queue for a message that matches your correlation ID.
In MQ you can listen for a reply in two ways, Listen and Wait, or Listen No Wait.
Now I have another vendor product that needs to obtain its Name/Address information from our SOR, however it requires a web service interface for a real time Name/Address lookup.
This requiring a synchronous interface, I need an integration component between the two
Sync-over-Async is an anti-pattern from what I am reading and I can understand why
I was thinking that maybe we need to just listen for a reply and wait. Everybody here is saying "Oh stop worrying Dustin, this mainframe is ludicrously fast, it would be extraordinarily rare for it to be unresponsive."
So maybe we just agree to a good timeout period? Wait for 1 minute to get a reply and if nothing comes back in that time frame then log an exception and return an error to the vendor software.
Problem is that if that message shows up an hour later in the reply queue then nothing is listening for it and it will just sit there
If this happens often enough then some poor schmuck has to be an MQ janitor now. I have seen developers fall into a support role on MQ based applications where they are little more than glorified janitors
So yeah there really is no perfect solution here
 
ratchet freak
ratchet freak
11:19 AM
set a listener deamon that listens for timed out request just to clean them up
 
 
2 hours later…
rolfl
rolfl
1:15 PM
Ba-dum-tish... Would this question be on-topic on Programmers? codereview.stackexchange.com/q/64531/31503
 
ratchet freak
ratchet freak
seems to simplistic
but honestly trying to merge the transitions of different automata is flawed IMO
 
rolfl
rolfl
That would make a good answer here on Programmers, right?
How about you prepare your FGITW for an imminent migration?
 
ratchet freak
ratchet freak
I would VtC it as it stands
 
Thomas Owens
Thomas Owens
@ratchetfreak Really?
It seems like a legitimate question to me...
 
rolfl
rolfl
There's little doubt about the status on CR, and it seems reasonable enough for me to run it through the Programmers gauntlet.
0
Q: Class template specialization for implementing Automata

TimI'm implementing classes for simulating and generating different kind of automata. I'd prefer to use the same State and Transition classes for all the automata: NFA, DFA, PDA, etc. For a PDA a transition from one state to another work require some sort of an extension, because the transition req...

java
Will need some retagging, etc.
You have no automata ? hmmmm
 
Thomas Owens
Thomas Owens
1:33 PM
@rolfl That's probable. Automata are a CS thing. But it's more of a software design question, not about the theoretical construct.
 
 
1 hour later…
Ampt
Ampt
2:36 PM
I had to go look automata up
 
user41796
3:09 PM
@Ampt I went "blah, blah, blah, whatever...." :-D
 
Jimmy Hoffa
Jimmy Hoffa
@maple_shaft yeah, this explanation makes far more sense :) It sounds like you have a very firm grasp on the situation, only thing I can mention is the same thing I mentioned to @GlenH7 the other day- in message based systems the idea of a supervisor is common which is notified of messages being sent around and keeps an eye on the workers dealing with them to ensure TTLs, handle cleanup when failures occur
grabs messages that are unknown/unexpected and troughs them off somewhere to be handled by the archival queue which has a longer memory perhaps et al
 
user41796
@JimmyHoffa The dead letter queue is a pretty common way of handling that in MQ systems.
 
user41796
And then some poor shmoe has to evaluate those periodically
 
Jimmy Hoffa
Jimmy Hoffa
@GlenH7 yes but the DLQ tends to be associated with manual work- a supervisor is an automated failure handler that maintains a working knowledge of all the stuff that has happened in the system so it can come up with an appropriate action on it's own rather than requiring some poor shmoe to dig through logs to find out what should happen to that message and do it
 
user41796
@JimmyHoffa From my MQ experience, failures and resulting DLQ messages are typically rare. So for me, automating handling of those hasn't provided enough benefit to justify the cost.
 
Jimmy Hoffa
Jimmy Hoffa
3:16 PM
the key to the supervisor is it monitors active work and has a long memory, perhaps in a DB or whatever, but it logs the relevant information about work being done to know what to do in the various failure scenarios that could occur- basically the info a manual operator would have to go dig up on their own to make the decision: Has this message already been received and handled? Is this message a response to work that was resubmitted and handled already? etc
 
user41796
Typically, the DLQ sits empty until a programmer makes a change. Then it fills up, they figure out what they did wrong, and then it empties back out.
 
Jimmy Hoffa
Jimmy Hoffa
@GlenH7 yeah, often that's the case, but he's working with SoR banking information, it can be pretty critical that nothing slips through the cracks (bank misses a name change for someone and it can misreport to the IRS or who knows what)
aye, it may be unnecessary, the majority of the time it's true to go with:
4 hours ago, by maple_shaft
I was thinking that maybe we need to just listen for a reply and wait. Everybody here is saying "Oh stop worrying Dustin, this mainframe is ludicrously fast, it would be extraordinarily rare for it to be unresponsive."
 
user41796
I got the impression all of that came through the batch and real-time system interfaces, so MQ wasn't going to be involved with those.
 
user41796
But if everything was to go through MQ.... Bleh. queues have additional challenges with transactional changes
 
Jimmy Hoffa
Jimmy Hoffa
@GlenH7 he's saying MQ is the only interface, and he's trying to design batch and real-time systems to work on top of that
(that's how I understand his trouble?)
 
user41796
3:25 PM
Nah, go look at the diagram again. He shows 4, 5? interfaces to the data layer
 
Jimmy Hoffa
Jimmy Hoffa
4 hours ago, by maple_shaft
The only means I have to talk to this world is through this MQ interface which is asynchronous. The model here is how one would typically do request-reply with MQ. Send messages to the request queue with a correllation ID, then have another process/thread listening on the reply queue for a message that matches your correlation ID.
 
user41796
OK, maybe I misread then. Or maybe what he wrote here doesn't align with what I remember of the question he posted
 
Jimmy Hoffa
Jimmy Hoffa
his Q wasn't super clear on why he was having an issue, scroll up, he described a genuinely problematic scenario above
 
user41796
I think we'll need him to drop back in here and explain then. I see what you're saying, but I'm not positive that's the case. Here's his question.
 
user41796
I think the Batch Job is just that - after hours processes that handle rectification between bank + fed + others + whatever.
The real-time synchronous interface is the teller system + whatever the other bank folk use.
 
user41796
3:30 PM
I think he's responsible for the bottom "Another widget" and maybe the interface to the "third party widget"
 
user41796
But I definitely could be misreading things
 
user41796
3:43 PM
@enderland Sad Trombone for you...
 
enderland
enderland
I had higher hopes than what I think will ultimately come true for this process :(
@GlenH7 that interview for the IT position went well, seems like a good possibility...
 
user41796
Maslov's hierarchy of needs...
 
enderland
enderland
It's likely a good stepping stone into the big data stuff at my megacorp too, plus allows working remote/from home and has a flexible work time schedule - seems legit
only thing is I think I'm way more qualified than the manager expects, heh
 
user41796
@enderland That's not always a bad thing if they can tolerate the risk that you'll move on after some period of time.
 
enderland
enderland
Yeah. We had a very frank convo which basically was him wanting me to know that the job might be slightly repetitive, to which I basically replied, "that's fine - just expect me to try to improve it"
Sounds like it'll have relatively low expectations, so lots of autonomy, that to me is a good situation
 
user41796
3:56 PM
Might not get you promoted, but sounds like it could certainly be used as a strong stepping stone.
 
enderland
enderland
Yeah, though autonomy is more valuable than a promotion in many regards anyways
 
user41796
For the path you're wanting to follow, definitely
 
enderland
enderland
It feels like cheating when you are prepared for an interview, though
 
user41796
Nah, not at all. That's coming from the mindset of interviews being adversarial. When you think of them as collaborations your perspective shifts dramatically.
 
enderland
enderland
I wrote out a bunch of STAR interview answers which I knew were going to directly map to the questions they would ask me (our internal process is based on our competencies for each position) so it was trivial for me to have good answers
 
user41796
4:05 PM
So they're trying to give internal candidates an edge... :-)
 
user41796
Oh, to go to the corporate holiday party or to not go....
 
user41796
Generally I don't. Nor does my boss. My boss' boss does, but only because he's expected to due to his role.
 
Robert Harvey
Robert Harvey
4:20 PM
Is there any benefit at all to learning C before C++?
I'm asking here because I already know what the folks in The Lounge are going to say.
 
user41796
@RobertHarvey what domain?
 
user41796
previous one you've been looking at?
 
Robert Harvey
Robert Harvey
Uh, embedded systems in aerospace?
 
Ampt
Ampt
yes
absolutely
not all embedded systems support C++
 
user41796
C++ is very much a newbie in the embedded arena
 
user41796
4:22 PM
most legacy stuff, aka what you're likely to deal with, is going to be C
 
Jimmy Hoffa
Jimmy Hoffa
I don't know enough about C++ to say this with authority, but I feel comfortable regurgitating this bit that I've heard endlessly from C++ folks: The vast majority of C++ is actually just C with classes - so given that, learning C will get you most of the way into being a productive C++ developer it sounds like.
 
Robert Harvey
Robert Harvey
Really? Maybe Northrop is making the move. None of their job openings mention C, only C++.
 
user41796
They may have had problems hiring for just C
 
Jimmy Hoffa
Jimmy Hoffa
most C++ folks can't agree on what the good parts of C++ are or what the right way to do any given thing in C++ is, which is largely why I think folks fall back to just writing C in C++
 
user41796
C++ gives you OO constructs which simplify maintenance of code
 
Jimmy Hoffa
Jimmy Hoffa
4:24 PM
@RobertHarvey most job descriptions mention stuff they wish to be doing- any C shop will have managers convinced they should be writing C++ because that's the improvement over C, right?
 
user41796
C is a much smaller space to learn as well, so there's a lot to be said for that
 
Thomas Owens
Thomas Owens
@RobertHarvey Are you coming to the aerospace side?
 
ratchet freak
ratchet freak
if you forget templates then C++ is pretty simple
 
Ampt
Ampt
@ThomasOwens He is joining the dark side.
 
ratchet freak
ratchet freak
but once you try anything non trivial with them they start to break down
 
Robert Harvey
Robert Harvey
4:25 PM
@ThomasOwens I worked at Armstrong Flight Research Center for 7 years, so I've been in aerospace for awhile.
 
Jimmy Hoffa
Jimmy Hoffa
@ratchetfreak but without templates, don't you basically stop yourself from using any of the libraries everyone claims to be the only way you should be doing anything?
 
Thomas Owens
Thomas Owens
@RobertHarvey Oh. I thought you worked for some financial thing.
 
Robert Harvey
Robert Harvey
That was Employer^^.
@JimmyHoffa You don't need to know template programming to use templates written by others.
Just the syntax for calling them.
I like the idea of learning C++ as C with classes. Poring through a 1000 page book on C++, not so much.
 
ratchet freak
ratchet freak
@RobertHarvey exactly
 
Robert Harvey
Robert Harvey
Ah, well the syntax for calling them is not nearly as complicated as writing them and reasoning about them.
 
user41796
4:28 PM
And double check what C standards each compiler supports.
 
ratchet freak
ratchet freak
trying to create a templated anything that complies with best practice is a minefield
 
user41796
I think C99 gets partial support. Don't think many / any have implemented C11 changes.
 
user41796
C++11 is pretty well supported by most of the compilers now though
 
user41796
and C++11 != C11
 
Jimmy Hoffa
Jimmy Hoffa
@ratchetfreak is that true though? When I tried doing some C++ recently, templates were literally everything, and while I didn't need to know them, I found it pretty confusing trying to figure out how I was supposed to compose them together without understanding their underlying semantics
 
Ampt
Ampt
4:29 PM
Personally I think you need to learn assembly before moving to C
 
ratchet freak
ratchet freak
you'd think writing a iterator would be easy, but no there has to be 5 different specialised templates for them
 
Robert Harvey
Robert Harvey
Why? I already know some assembly, but programming in assembly is like trying to juggle Wile E. Coyote bombs.
 
Jimmy Hoffa
Jimmy Hoffa
@ratchetfreak exactly my point- would understanding templates make this clearer? Or is it just as ugly as me not understanding templates and you just kind of guess them together until they work?
 
Ampt
Ampt
@RobertHarvey In Soviet Russia, only program assembly. C for kapitalist dog.
 
user41796
@RobertHarvey because he obsesses over picosecond efficiencies.
 
Jimmy Hoffa
Jimmy Hoffa
4:31 PM
@GlenH7 and still finds the ground fell out from under him
 
Robert Harvey
Robert Harvey
When I owned a 286, I had a Russian copy of Welltris. It was capable of producing polyphonic music through the internal speaker. I was impressed.
Pretty sure it was written in assembly.
 
Ampt
Ampt
@JimmyHoffa Yeah but I polled the distance to the ground at least 3 million times per second while it fell.
 
user41796
Likely yes. Crazy strong math and logic skills in the former USSR
 
Jimmy Hoffa
Jimmy Hoffa
@GlenH7 truth. The russians I worked with were absolute C/C++ wizards, circa '08 (they should have been beyond this)
 
user41796
Typically a very deep understanding of the underlying architecture and what the hardware was doing.
 
ratchet freak
ratchet freak
4:35 PM
@JimmyHoffa C++ templates can do with a from-scratch redesign
 
Robert Harvey
Robert Harvey
Embedded, real-time programming has significant constraints. I found a coding standard a few months ago for C from one of the aerospace companies. It forbids things like memory allocations in loops and recursive function calls.
OK. Well, I guess I'll either find a "C++ is just C with classes" book, or get started on Learn C the Hard Way.
 
user41796
@RobertHarvey I think you'd be safe starting with C as your foundation. K&R is dated but it still gives a really good overview of what the language was intended to do
 
user41796
And be careful with books from Herbert Schildt on C. The ones I have seen have had a number of errors in them.
 
ratchet freak
ratchet freak
I went from java to C++ and I generally stay away from templates except for the most basic stuff
 
user41796
@RobertHarvey I want to say I have heard of Richard Barnett outside of the C embedded space, but it looks like his books may be a good start.
 
Thomas Owens
Thomas Owens
4:48 PM
What, exactly, do you mean by "embedded", though.
 
user41796
In the above case, dedicated PLCs. Although I don't think PLC is the right general term
 
user41796
But as the cost of chips plummets, you can get more and more functionality out of that chip.
 
enderland
enderland
@GlenH7 PLCs are more specifically for controlling manufacturing processes
 
Thomas Owens
Thomas Owens
Where I am now, our definition of embedded is usually a single board computer running Linux or vxWorks and having at least a gig of onboard memory, usually that interact with hardware components directly over physical connections.
 
user41796
From my naive perspective, that seems to be on the big end of embedded. Would exclude a lot of smaller chips where you're still dropping a compiled assembly onto them.
 
Robert Harvey
Robert Harvey
4:58 PM
 
user41796
Yeah, I think I need some new toys too. :-)
 
Robert Harvey
Robert Harvey
Fifty five bucks for a Linux board, 1Ghz, 512MB SDRAM, 8GB flash.
Or, if you want to go smaller, the Tiva C evaluation board for 14 bucks.
32 bit, 80Mhz, 32K SDRAM, 2K EEPROM. USB interface.
Has the prettiest Assembly Language I've ever seen.
 
user41796
Blasphemer. We all know that title goes to the Motorola 68* series... :-)
 
Robert Harvey
Robert Harvey
5:13 PM
@GlenH7 Yes, this looks very similar to the AL that TI is using.
 
user41796
IIRC, Motorola made the chip available to universities for ridiculously cheap. Combined with a flood of instructional material on it, it was the de facto choice for years for a chip to learn on.
 
Robert Harvey
Robert Harvey
Hmm... Motorola embedded boards seem hard to find.
 
user41796
They have really fallen out of favor. There are cheaper, more powerful chips available. Most have languages or constructs similar to what you'd see with "bigger" chips.
 
Robert Harvey
Robert Harvey
OK, that's a little creepy. Because when I think "frosty woman," I think embedded systems.
 
Jimmy Hoffa
Jimmy Hoffa
@RobertHarvey really? because it just reminds me of Doctor Who...
 
enderland
enderland
5:31 PM
@RobertHarvey something I did for my last interview which was helpful, and might be relevant to you and @MichaelT is actually writing out STAR interview responses for a variety of situations, was pretty useful and might be of interest
 
Robert Harvey
Robert Harvey
OK. I'll look into it.
 
enderland
enderland
we generally tend to focus way too much on background info for all examples we talk about, rather than what we did and what resulted from it (which is more what people actually care about)
 
harish.venkat
harish.venkat
0
Q: Java Program High CPU usage

harish.venkatI wrote a java program which queries oracle database and converts it to object using beanutils During Load Testing I found high CPU usage. Outline of the program is: List<Map<String, List>> rows = queryList(query, new Object[]{params..}); while(rows.hasNext()){ beanutils.setProperty(obj, ...

java performance jdbc cpu-usage apache-commons-beanutils
does these questions get answered here or I am at wrong place?
 
Robert Harvey
Robert Harvey
@harish.venkat This looks like your typical "Hail-Mary" question. "I have no idea what's going on so I'm going to throw it out there in the hopes that someone has seen my exact problem before."
 
harish.venkat
harish.venkat
@RobertHarvey Probably but I need some help on how to proceed & I hope this forum does not forbid it
 
Robert Harvey
Robert Harvey
5:46 PM
@harish.venkat That BeanUtils call looks like some form of Reflection. Is it?
 
harish.venkat
harish.venkat
yes that uses reflection
 
Jimmy Hoffa
Jimmy Hoffa
@harish.venkat not a forum, this is a Q&A site focussed on answering narrow questions to collect authoritative answers with a similar goal of Wikipedia wanting authoritatively correct information
 
Ampt
Ampt
Helping people is a byproduct really
@harish.venkat reflection is really, really performance heavy
well, it can be, depending on how it's implemented
it's a fairly advanced concept, I suggest you start by understanding what it's doing and move forward from there
 
Jimmy Hoffa
Jimmy Hoffa
reading the question in detail, it's not half bad
 
Robert Harvey
Robert Harvey
@harish.venkat Have a look at these performance graphs‌​. Apache Commons Beans is an order of magnitude slower than some other techniques.
 
Jimmy Hoffa
Jimmy Hoffa
5:50 PM
maybe some edits could help? But I think it's actually not a terrible question offhand... it has a lot of details which really helps
 
harish.venkat
harish.venkat
@JimmyHoffa thank you
does waiting on query takes cpu utilization on the host?
 
Robert Harvey
Robert Harvey
Reflection is CPU intensive, if that's what you mean.
 
harish.venkat
harish.venkat
yes I got reflection is cpu intensive but my concern is quite general..see the last part of my question
 
Jimmy Hoffa
Jimmy Hoffa
@harish.venkat that can vary as to what the underlying wait is doing- is it look like a "wait" to you but it's doing a bunch of parsing in the library below you? Is it waiting on the OS to do the actual network IO while you have a software networking driver spiking your CPU handling the framebuffers because of low-quality networking hardware/drivers?
Lots of things could be spiking your cpu below your wait
Reflection isn't always CPU intensive, it's more IO intensive in .NET interestingly (it has to go back to the assembly's IL repeatedly, it becomse CPU intensive when you start doing invocations that cause a lot of JITing)
but that's .NET, reflection is an advanced thing with many approaches- it could be the cause
 
Robert Harvey
Robert Harvey
@JimmyHoffa How big are assemblies? Surely the OS is caching/buffering that sort of thing.
 
Jimmy Hoffa
Jimmy Hoffa
5:55 PM
I think what you really need to do @harish.venkat is instrument your code. Put in stopwatches, measure, measure, and measure some more. isolate each little piece and figure out precisely which call is causing your CPU usage to spike. You may want to find an instrumentation library somewhere that can hand you CPU usage metrics at points in time or just use an actual profiler
 
Robert Harvey
Robert Harvey
He kinda already did that. He replaced the Beans call with a simple Getter/Setter, and the problem went away.
 
Jimmy Hoffa
Jimmy Hoffa
@RobertHarvey the fusion resolution isn't cached because the available assemblies in resolution path may have changed, so it has to do some folder/file grepping to find what you asked for each time, like I said it can be CPU intensive depending on exactly the reflection actions you're taking
@RobertHarvey ...then what's his question? It looked like he was asking which line is causing so much CPU utilization
 
Robert Harvey
Robert Harvey
Yeah, we already know that.
Really, the question amounts to "How do I get this expensive call to an outside library to be cheaper?" And the answer is: either change the code in the library, or use something else.
 
Jimmy Hoffa
Jimmy Hoffa
reflection can be IO intensive, it depends on the rate you're doing it and the actions you're taking. Reflection invocation isn't the majority of reflection people do, but it is more resource intensive than other activities (inspecting types trying to find particular things et al, this is a lot of asking the IL for offsets)
@RobertHarvey that's not how I read it at all...
perhaps edit it as such?
It seems like a clear "What is causing my CPU to spike? This? That? Which?"
 
Robert Harvey
Robert Harvey
Yes, but the answer is already in his question.
 
harish.venkat
harish.venkat
6:01 PM
No my doubt is
 
Jimmy Hoffa
Jimmy Hoffa
@RobertHarvey I read that as speculation
 
harish.venkat
harish.venkat
is it really the reflection causing the problem or my multithreading causing the problem
 
Robert Harvey
Robert Harvey
Well, that a getter/setter fixes the problem is a fairly conclusive experiment.
You solve the problem by removing the Bean call.
 
Jimmy Hoffa
Jimmy Hoffa
@RobertHarvey did it? I didn't understand it as such... again, perhaps the Q could use some editting- though it does appear to be a good Q
 
harish.venkat
harish.venkat
my threadpool size is 8..so threads will wait on the monitor and after the query is finished it will execute the mapper step
 
Robert Harvey
Robert Harvey
6:03 PM
@harish.venkat: Did converting the Bean call to an ordinary Getter/Setter solve the CPU spike?
 
harish.venkat
harish.venkat
yes it solved but I did not get why cpu utilization is at 50% still
 
Thomas Owens
Thomas Owens
Why would one include architectural diagrams in requirements specifications?
 
Robert Harvey
Robert Harvey
@harish.venkat Leave the getter/setter in, and run a profiler on it.
 
Jimmy Hoffa
Jimmy Hoffa
@ThomasOwens micromanagement- those dumbass coders don't know what they're doing, they'll fuck it up, I need to tell them exactly how and what to do.
 
harish.venkat
harish.venkat
and my concern is later my computation in my api increases should I decrease my cpu threads
 
Thomas Owens
Thomas Owens
6:04 PM
@JimmyHoffa I...what?
 
Robert Harvey
Robert Harvey
@ThomasOwens Requirements, or Specifications?
 
harish.venkat
harish.venkat
*threads accessing the api or should I change the db threads
 
Robert Harvey
Robert Harvey
@harish.venkat You're guessing. Run a profiler on it. It's the only way to be sure.
@ThomasOwens Jimmy forgot to take his lithium this morning.
 
Jimmy Hoffa
Jimmy Hoffa
measure, measure, and measure some more. Don't ever guess about performance stuff.
3
 
Thomas Owens
Thomas Owens
@RobertHarvey Not this. Requirements Specifications are different than Design Specifications.
 
Jimmy Hoffa
Jimmy Hoffa
6:07 PM
@ThomasOwens ? doesn't make sense?
 
Thomas Owens
Thomas Owens
@JimmyHoffa Not really. I don't even know how to create an architecture without requirements.
 
Robert Harvey
Robert Harvey
Yeah, doesn't belong there.
That would be a neat trick.
 
Jimmy Hoffa
Jimmy Hoffa
I think I must be misunderstanding something about your scenario....where'd I put that lithium again...
 
harish.venkat
harish.venkat
ok If I have 20 threads that accepts request for this api...I have restriction on db as 8 threads
By this structure of query the db and do the mapper:
I can have 8 threads running the query and another 12 threads running the mapper which on combination causes the cpu to spike.
I have tried to put my scenario more appropriately
 
Robert Harvey
Robert Harvey
Use. A. Profiler.
 
Jimmy Hoffa
Jimmy Hoffa
6:12 PM
@RobertHarvey big fan of the UAP technique
 
Robert Harvey
Robert Harvey
 
harish.venkat
harish.venkat
thanks, will update the question after doing more profiling
 
TGMCians
TGMCians
hi guys
 
Robert Harvey
Robert Harvey
Well, hopefully the profiler answers your question. There isn't anything else anyone can tell you that the profiler won't.
TGMCians in da house.
 
TGMCians
TGMCians
yo
 
Ampt
Ampt
6:16 PM
WUUB WUUB WUUB
 
Robert Harvey
Robert Harvey
[makes turntable scratching noises]
 
Jimmy Hoffa
Jimmy Hoffa
@Ampt damnit use the ventilation system and shut that window!
 
Ampt
Ampt
@JimmyHoffa WUBWUBWUBWUBWUBWUB
 
Jimmy Hoffa
Jimmy Hoffa
@Ampt no worries, I just got all 4 replaced this week and a full lifetime warranty on them so I can replace the one with a brand new one and it won't have tread different from it's opposite
 
Robert Harvey
Robert Harvey
6:42 PM
Well, The Lounge is claiming that all of the tools in the C++ tool shed have no protective guards, and that you have to be proficient at all of them to use the language.
 
psr
psr
Also to be awesome.
 
Robert Harvey
Robert Harvey
I couldn't pin them on a functional subset.
"Nope. You need to learn all of it. Through blood, sweat and tears."
 
psr
psr
Which means that they're basically arguing that C++ is a lousy language.
 
amon
amon
“functional subset”? Sure, C++11 has lambdas. You can do ordinary FP with that, although memory management will get hairy.
 
Robert Harvey
Robert Harvey
"Serviceable subset."
The 20% of the language (and libraries) that everyone uses. Apparently C++ is the one language/platform that requires you to know all of it.
 
Matt Giltaji
Matt Giltaji
6:52 PM
yay for the STL :D
 
Robert Harvey
Robert Harvey
in Lounge<C++> on Stack Overflow Chat, 21 secs ago, by Jefffrey
By this time you probably discover Haskell, and just forget about C++ entirely.
 
user55340
I'm curious where the todo database schema was advertised... I've got 29 up votes on it today
 
user55340
47
A: Database schema for a ToDo list

MichaelTThere's a joke I heard awhile back: Q How does a BASIC coder count to 10? A 1,2,3,4,5,6,7,8,9,10 Q How does a C coder count to 10? A 0,1,2,3,4,5,6,7,8,9 Q How does a DBA count to 10? A 0,1,many The truth behind this joke is that once you have two (or more) of the same thing ...

 
Thomas Owens
Thomas Owens
7:15 PM
It was tweeted 23 hours ago.
Database schema for a ToDo list http://programmers.stackexchange.com/q/261269?atw=1 #database
Tweeted by StackProgrammer on October 29, 2014 at 8:04 PM
 
user55340
@ThomasOwens Lots of questions are tweeted.
 
Thomas Owens
Thomas Owens
7:57 PM
I don't understand the concept of "high level" and "low level" software requirements. It apparently comes from one of the aerospace standards, but I've never heard of anyone asking for it.
I should read this thing from the FAA, I guess.
 
Ampt
Ampt
Startrek-TNG Season 2 Episode 22: Because we need a recap of every situation in which Riker got some.
 
Thomas Owens
Thomas Owens
@Ampt Which episode was that?
 
user55340
That was actually kind of fun to dig into...
 
user55340
0
A: Floating point number to binary

MichaelTThe first thing to realize is that the binary representation of 0.15625 is not 0.00101. Yes, that is what you would write if you were writing out the number by hand. The actual representation of the number within the computer using IEE 754 for single precision 32 bit. This is specified as part...

 
Ampt
Ampt
@ThomasOwens Riker gets alien virus and they have to make him dream to save him. The first half of the episode is him dreaming about all his conquests from the series so far
@MichaelT Ugh, Fine, have an upvote
 
user55340
8:08 PM
@Ampt You could have waited 4 hours...
 
Ampt
Ampt
@MichaelT HA!
 
user55340
Shades of Gray (Star Trek: The Next Generation)
"Shades of Gray" is the 22nd episode of the second season of the American science fiction television series Star Trek: The Next Generation, the 48th episode overall. It was originally broadcast on July 17, 1989 in broadcast syndication. It was the only clip show filmed during the series and was created due to a lack of funds left over from other episodes during the season. Set in the 24th century, the series follows the adventures of the crew of the Federation starship Enterprise. In this episode, Commander William Riker (Jonathan Frakes) undergoes medical treatment by Dr. Katherine Pulaski (Diana...
 
Ampt
Ampt
should have waited 4 hours to post it
 
user55340
(and the memory alpha site: en.memory-alpha.org/wiki/Shades_of_Gray )
 
Thomas Owens
Thomas Owens
Oh.
Yeah.
 
user55340
8:10 PM
> This treatment finally eradicates the infection, and Riker recovers to his well-adjusted, humorous self.
 
Ampt
Ampt
@MichaelT DUDE!! SPOILERS!!
 
user55340
@Ampt But... only 5 more answers until silver java tag badge.
 
user55340
8:34 PM
 
user55340
>
Human collective behavior can vary from calm to panicked depending on social context. Using videos publicly available online, we study the highly energized collective motion of attendees at heavy metal concerts. We find these extreme social gatherings generate similarly extreme behaviors: a disordered gas-like state called a mosh pit and an ordered vortex-like state called a circle pit. Both phenomena are reproduced in flocking simulations demonstrating that human collective behavior is consistent with the predictions of simplified models.
 
gnat
gnat
@MichaelT I think you are my hero
 
Jimmy Hoffa
Jimmy Hoffa
just such an awesome exchange about SOAP, @MichaelT having lived through all of this you should really appreciate this:
 
Robert Harvey
Robert Harvey
9:11 PM
Y'know, you would think that after all of these years certain things would be easier to do on computers than they actually are. Like restoring file associations in Media Player. Or sorting a list of files in Windows Explorer by something other than name.
 
gnat
gnat
9:28 PM
those who can't make a good career question, dump their attempts at Programmers. I hate Agile, how to avoid it in my career. Those who can't make a good programmers question, dump at Workplace. I want to build my career, how to proceed if I hate Agile
 
psr
psr
questions about hating agile are on topic everywhere.
 
gnat
gnat
@psr try it at Code Golf?
On top of that, for experienced programmers all the SCRUM rituals might just get in the way of getting thinks done. To continue with your metaphor: if you are driving a Ferrari on a straight road, having to stop every 2 km or so to check if you are going in the right direction will only make you slower. But, yes, it will help (bad) managers to have a feeling of control. — Giorgio Aug 24 at 13:39
 
Ampt
Ampt
9:55 PM
@psr psh, speak for yourself. I enjoy Agile.
 
Matt Giltaji
Matt Giltaji
@Ampt we named my company's implementation of it Fragile
 
Ampt
Ampt
You guys gotta be doing it wrong
 
Matt Giltaji
Matt Giltaji
@Ampt yeah we did
 
Ampt
Ampt
The only place where it sucked was at school. the two non-academic implementations I've used since have been great
 
Matt Giltaji
Matt Giltaji
they looked at the agile manifesto, "adopted" it, then promptly declared half of it as "not applicable to our business situation"
 
psr
psr
9:59 PM
@Ampt I've pretty much experienced scrum as micromanagement. It's pretty much used to make sure programmers don't spend time on tasks that management doesn't understand, even if they are necessary for the stories.
 
Matt Giltaji
Matt Giltaji
no one had much experience with it, including the "scrummasters", but they felt that it was workable with 3 teams of about 30 people each
 
Ampt
Ampt
@MattGiltaji that's too big
 
Matt Giltaji
Matt Giltaji
@Ampt ya, we told them to start small
but they wanted to get the project done faster
each user story had a mini-waterfall process associated with it
 
psr
psr
Also, @Ampt, if scrum is about self organizing and empowering, how come the thing it's best known for, and named for, the stand-up, is based on management controlling the physical position of your own body?
 
Ampt
Ampt
@psr Your scrum-master has to A) be technical, and B) Be capable of persuiding the product owners (management, stakeholder, whoever) about the technical merits of doing that kind of stuff
@psr You've clearly already made up and closed your mind on this topic. I don't think that this conversation is going to be productive for either of us.
 
Matt Giltaji
Matt Giltaji
10:01 PM
and they went through about 3 sprints without getting a single story done
i would love to do agile at a place that does it properly
 
psr
psr
@Ampt But you don't get to do the "No True Scotsman" thing. You have to count bad Agile as Agile. Any method is good if you do it in a successful way.
 
Ampt
Ampt
having the tasks under a story be sequential isn't too uncommon. you gotta design and implement before you can test, no?
 
psr
psr
@Ampt Closed as not an answer.
 
gnat
gnat
@MattGiltaji half-arsed
Individuals and interactions over processes and tools
and we have mandatory processes and tools to control how those individuals (we prefer the term ‘resources’) interact

Working software over comprehensive documentation
as long as that software is comprehensively documented

Customer collaboration over contract negotiation
within the boundaries of strict contracts, of course, and subject to rigorous change control

Responding to change over following a plan
provided a detailed plan is in place to respond to the change, and it is followed precisely
2
 
Ampt
Ampt
@psr Actually, I do. Look at the Agile Manifesto
Individuals and interactions over processes and tools
it's right on the cover
if you're putting processes over individuals and interactions, you, by definition, aren't doing agile.
 
Matt Giltaji
Matt Giltaji
10:04 PM
@Ampt this was more of a process over individual kind of thing
 
Ampt
Ampt
 
psr
psr
@Ampt But if most of the people that say they are doing Agile put process over individuals and interactions, that's not really what Agile is. But yes, I'm in favor of Agile that only does good things, successfully.
 
Ampt
Ampt
@psr what you've done is attempted agile. It sounds like you never succeeded. As to your point about standing, the idea is that the developers are in a position that isn't comfortable, so that they don't dwell on small problems at the standup. the standup is for reporting, not problem solving
that's why you stand
 
psr
psr
@Ampt The only time I've seen these things done, people didn't call it agile.
@Ampt I know the reason. It simply isn't consistent with self-organizing and empowering.
It's dictating a solution to the problem
 
Ampt
Ampt
Where are you getting that Agile must be "self-organizing and empowering"?
where is that listed as a goal?
@psr It's empowering for the developers who's time would be wasted by solving problems in a meeting they don't need to be in.
 
psr
psr
10:10 PM
@Ampt No, it's just helpful. Empowering would be allowing them to leave.
 
Ampt
Ampt
8 mins ago, by Ampt
@psr You've clearly already made up and closed your mind on this topic. I don't think that this conversation is going to be productive for either of us.
 
Matt Giltaji
Matt Giltaji
I would say that sprint planning/retros and doing story points are fairly empowering
(until they start getting tweaked by PMs after the fact because "the velocity isn't high enough")
 
psr
psr
@Ampt Again, I'm actually for Agile as the manifesto says it. I just think those principles have a slight negative correlation to people saying they are practicing Scrum, and zero correlation to people saying they are practicing Agile. Could be I've seen a biased sample.
 
Ampt
Ampt
@MattGiltaji Yeah, you've really got to have a "safe" zone for the numbers and metrics. If people think that they are being judged on the metrics they will, even if unconsiously, change how they estimate
"We aren't meeting managements velocity? All our estimates are now 1.5x what they were"
but if you do have that safety, you can get some pretty damn accurate estimates
 
Matt Giltaji
Matt Giltaji
like I said, we did pretty much every part of agile wrong
 
Ampt
Ampt
10:14 PM
Hey, but at least you know that, right?
 
Matt Giltaji
Matt Giltaji
i think almost everyone knew that except the people in charge
 
Ampt
Ampt
Yeah that sucks. Kind of ruins it for everyone unfortunately
It takes a lot of faith from management to give up that kind of control though. Probably why it ended up not working. it requires a large change
 
Matt Giltaji
Matt Giltaji
I think it could have worked, had we started with a small core group that worked well together, then expanded after they were comfortable with it
 
Ampt
Ampt
Yeah, the first place I worked with it had 6 devs and a manager who was our scrum master
he was great at talking technical to management
probable because he wasn't technical himself, but was willing to listen and learn from our concerns
 
Jimmy Hoffa
Jimmy Hoffa
@Ampt sadly many claim Scrum to be different from Agile- it's not the Scrum manifesto right? I wish I could use the agile manifesto to shutup some of the shitty scrum implementations I've seen...
(I have seen good scrum too)
@Ampt sounds like you basically had a good manager, I wouldn't blame scrum for the successes of good management
 
Ampt
Ampt
10:22 PM
Ya'll are some cynical bastards some days. You know that right?
 
Jimmy Hoffa
Jimmy Hoffa
true- one thing that surely happens plenty is places are doing a shit job of producing, everyone knows it, so they go for this scrum thing to try and fix it, trouble is they have shitty management and they screwed everybody's productivity and quality up before scrum, during scrum, and after scrum
@Ampt let the industry stew in you for a while :P We're actually surprisingly upbeat about a lot of things much of the industry is worse regarding.
we still largely enjoy tech- many in the industry abhor it
 
Matt Giltaji
Matt Giltaji
if only there was a way we could get paid to work with technology but not have to deal with any people
 
Jimmy Hoffa
Jimmy Hoffa
@MattGiltaji I'm not that cynical yet (I have a couple colleagues who are...)
y'know what, screw all that- @Ampt don't listen to us cynical bastards; keep that starry eyed optimism and you'll be management by 30, it'll be so much better that way.
(just don't be a dick to the engineers who work for you)
 
psr
psr
10:44 PM
It was weird, on that question Gnat linked about agile, how many of the answers talked about developers slacking off until a deadline loomed. I don't do that, and only a small % of people I've worked with do that. I have not told mgmt I'm done before a deadline, so I could use the time to do the important things they'd denied (and thank God I did).
Of course, that would look like I slacked off until the deadline, and would perpetuate the lack of trust that might have been what led them to deny developers the time to do those things in the first place.
 
enderland
enderland
So it sounds like bad managers are implementing bad processes and getting expected results, while good managers implementing good processes get good results- is that a tl;dr of the last convo?
 
psr
psr
@enderland Can't be, it didn't mention "agile".
 
enderland
enderland
@psr Agile is not the primary part of your previous conversation, managers not really following an established process and then screwing it up happens in way more ways than Agile
 
psr
psr
"bad managers are implementing bad processes and getting expected results, while good managers implementing good processes get good results, agile agile scrum agile" would be acceptable.
@enderland Hmm, some of what I'm talking about is managers establishing and following a process that embodies screwing it up.
They thought that process was "agile", though not all would agree.
 
enderland
enderland
Almost all managers hate the idea of "giving up control" and Agile doesn't exactly make them feel warm and fuzzy about that
 
psr
psr
10:50 PM
@enderland That's an interesting perspective. What do you think would be warm and fuzzy?
Or at least, more warm and fuzzy.
 
enderland
enderland
@psr Super predictable, waterfall like project progress they can meticulously track the project status against and report to their bosses the exact percentage of how on track projects are
 
user55340
11:47 PM
Strange things you find in Wikipedia:
 
user55340
Streaky the Supercat
Streaky the Supercat is a fictional character that appears in comic books published by DC Comics. Streaky first appeared in Action Comics #261 (February 1960) and was created by Jerry Siegel and Jim Mooney. == Fictional character biography == One of a series of super-powered animals including Krypto the Superdog, Comet the Superhorse, and Beppo the Supermonkey that was popular in DC's comics of the 1960s, Streaky was Supergirl's pet cat that was given super-powers by an unusual form of kryptonite. In Action Comics #261 (February 1960), Supergirl was experimenting on a piece of green kryptonite...
 
Robert Harvey
Robert Harvey
[Reads about Agile in The Whiteboard... Now knows even less about it]
 

« first day (1518 days earlier)      last day (3459 days later) »