« first day (1491 days earlier)      last day (3502 days later) » 
00:00 - 18:0018:00 - 00:00

12:29 AM
Damn you @MichaelT... you are not allowed to go around on SO, and answer questions i have answerred already, with answers that are so much better than mine....
Makes my answer look.... sparse.
And it makes me vote, which is not traditional on SO.
And I am off to grab a bite to eat... nice answer.
 
user55340
1:10 AM
@rolfl Blame my P.SE sensibilities of fully answering questions.
 
user55340
Though do note that I started writing it before I saw '1 answer' alert box show up.
 
user55340
1:23 AM
In random perusal of hot questions (judaism often shows up in there), saw an interesting advert there... to a meta question itself.
 
user55340
 
user55340
Then links to:
 
user55340
20
Q: Mi Yodeya Referencer: A Syntax for Linking to Sources

HodofHodSeveral times, the community has discussed and requested an easier way to link to sources. So recently, with the help of Hashem yisborach, and @Menachem, I threw together a quick script to do just that. It links to Tanach and most of Talmud Bavli, and with the help of the community, we've added ...

 
user55340
Not sure how we'd use it (most of our meta tends to be prescriptive rather than informative)... but maybe with the proper meta guidance for things.
 
user55340
1:27 AM
@RobertHarvey hey, you could use that to have a 'community advert' on SO that links to the close vote queue.
 
user55340
Or maybe a 'how to format your code' MSO question.
 
2:23 AM
@AshleyNunn fun with broken english: Wife asked me the other day if I had a good day, my response: "aye, did some chores 'rounna hahse 'nat" -> 2 fun language things I stole from living in pittsburgh: ow/ou->ah, downtown=dahntahn, "and all that" -> 'nat
shit like that is why I see my kid failing english... I wonder if "aye" for "yes" is even accepted as proper english? That one might be...maybe...
 
Hey all, visitor from TeX.SX over here. I'm a bit wary about asking a question I think may be off-topic here. I'm looking for a C++ reference manual in texinfo format (to be used within emacs)—would such a question be on-topic here?
 
@SeanAllred unfortunately no, resource requests aren't on topic, but asking stuff like that in chat is always fine - that said, not many C++ folks 'round here usually, you may want to hop into SO's chat rooms and find their C++ room (the lounge), they may be able to help you (inbetween being relatively inappropriate and beware of NSFW there...)
 
@JimmyHoffa Oh boy… heh. But thanks for the heads up :) I forgot that SO has a lot of tech-specific rooms :)
 
yep, no worries. They can be helpful, C++ folks just tend to have a very insular community and as such their culture can be a bit prickly (in my experience)
 
user15026
@JimmyHoffa I use it periodically, so I say yes
 
user15026
2:38 AM
@JimmyHoffa Also, trying to read this makes me curious how it sounds when you actually say it
 
I kind of drawl everything together a lot in my home speech... skip pronouns wherever I feel like it, refer to people with "oy" as replacement for 'hey you', my vernacular is sincerely broken
though it's nothing like a southern drawl, no twang in my voice.. anytime there's a "the" I tend to just make it an a at the end of whatever word came before, "inna" "rounna", etc
oh and I habit to use "vas" (was) from german when referencing a thing I'm inquiring regarding, "vas ist das" or just "vas" - my wife's used to it and doesn't notice because she knows what I mean...
the context is usually clear sometimes I use that outside of home and people usually just assume I stumbled over my words I think as they respond without really noticing
"did you see this?" "vas?" "there's a new horror movie out!" "oh aye"
 
 
11 hours later…
2:02 PM
Man people are sour about being told to not test their private methods
 
user55340
@Ampt Its one of those holy wars and misunderstandings.
 
Yeah, it definitely smells like one
All the zealots showed up to the party I guess
 
user55340
I don't go through hoops testing private methods, but I do give sufficient access to a lot of internals to the class... other things are indeed private, but can be fully exercised through public methods.
 
user55340
@Ampt Read The Way of Testivus and go get some beer and buffalo wings.
 
Took a look at some blog posts about #regions. I genuinely don't understand the hate. #regions are just another organizational tool; if your beef is with code folding, then why don't we just dispense with code folding altogether? — Robert Harvey 24 secs ago
 
user55340
2:06 PM
Btw, fun world news in sports... Oslo essentially said "screw you" to the IOCC leaving them with a choice of China or Kazakhstan for the winter olympics of 2022. cbsnews.com/news/…
 
user41796
@Ampt Not supposed to fiddle with your privates in public, right?
 
no one wants to host the olypmics? why
 
Why wouldn't you test something?
 
@GlenH7 I guess I'm just more forward thinking but I like to have my privates out where everyone can see them
 
user41796
2:07 PM
@MichaelT Cost to Benefit of being the host city has gotten way out of whack.
 
@MattS. I'm on the mindset that you should be able to thoroughly test your private methods from the public one
if you can't, you need to pull that functionality out
 
user55340
@Ampt No one wants to bribe the IOCC with another $50B after Sochi... and IOCC doesn't want to reuse existing facilities... like SLC or Vancouver. It takes decades to get the stadiums paid off.
 
@Ampt Well, that's just silly. You might as well just dispense with refactoring altogether.
 
user41796
@RobertHarvey Some people abuse them. Therefore they are evil and everyone should avoid using them.
 
2:09 PM
@RobertHarvey I don't follow
 
user55340
A picture from 2004 Athens stadium today...
 
user55340
 
The whole point of writing methods is to make a testable piece of code.
Doesn't necessarily mean you want it in your public API.
But it's still nice to be able to test it. For God's sake, is it really the position of the TDD'ers that "if it's not public, thou shalt not write unit tests against it?"
Ridiculous.
 
user55340
Didn't help that a newspaper released the demands that the IOCC had for Norway... vg.no/sport/ol-2022/… (norwegian)
 
user55340
> The high ranking IOC members must have a cocktail party with the King of Norway, paid for in full
Roads cleared for traffic and only to be used by high ranking IOC members during the stay
VIP lounges and bars with full service and staff 24/7
New cellphones for all the IOC members, Samsung brand
Dedicated cars with drivers for all IOC members and whoever else they see fit
Separate entries on the airport for high ranking IOC members
A ceremonious reception to be held at the air port for high ranking IOC members
 
user55340
2:12 PM
Its almost as if the IOCC wanted to give FIFA a run for their money as most corrupt organization.
 
@Ampt yeah, I was just thinking how people would catch mistakes without testing. But I guess people who don't test don't make mistakes...
 
What's up with everyone and this "Public API" crap. What are you trying to hide by making your methods private? Do you really think that not allowing people to call them is the best way to protect your class or keep things from getting out of order? We are talking about people who are coding against your class here.
@MattS. Who's not testing?
 
@Ampt Not testing private classes. Sorry, I'm a bit behind I think.
 
Writing a good API is hard. It's made harder by folks insisting that methods be made public that have no business being made public. Not ever bit of code you write is reusable.
 
user55340
@Ampt we're talking about your stupid cow-orker who will mess it up... or maybe the client who wants to maintain it and doesn't mind fiddling.
 
2:15 PM
So putting a private keyword on your method is going to fix your stupid co-workers or other folks from tarnishing your class? Right... I can see why we should litter our classes with extra junk to allow third classes to come in and test them.
That'll just make everything better
 
user55340
Hopefully it makes them think twice about doing it.
 
user55340
Once you go into making private methods public it means you can't refactor it because they are expecting that method to do something. And it means you need to maintain it.
 
Well if we're tying unit tests to all of our private methods we can't just go around deleting them either, right?
or changing the functionality drastically
 
user55340
You go through, you've got a private method. Someone makes it public and uses it. The class gets refactored so it doesn't use that method... now that method doesn't work anymore and something, somewhere you didn't even consider breaks because your object isn't changing state as expected.
 
> IOC members will be received with a smile on arrival at hotel.
 
user55340
2:18 PM
Making things private for things you shouldn't be calling reduces the long term maintenance work that needs to be done on the class as the innards get changed.
 
user55340
@MattS. did you see the length of the document? Those are just the highlights.
 
@MichaelT Oh man, I'm reading through it. I'm just wondering what type of person writes these things up? No wonder everyone is saying no.
 
user55340
@MattS. There's the "separate lanes / roads for IOC where ordinary people and public transport won't go"
 
user55340
> In total, the International Olympic Committee (IOC) 7000 pages of requirements and specifications to Oslo søkerby. Much is normal security requirements and arrangements, but many pages is also about how the IOC delegates will have the best OL, fired up with continuous access to alcohol, special service and luxury food.
 
user55340
2:22 PM
> * Traffic rules and traffic lights must be adjusted so that the Olympic traffic is prioritized
 
user55340
Remember London?
 
user15026
@MichaelT This list is crazy
 
user15026
I can totally understand why just about everyone is pulling out of this
 
> All meeting rooms must be exactly 20 degrees Celsius all the times.
 
This is nuts
 
2:24 PM
What did you do at the olympics? "I was the temperature checker."
 
user55340
@AshleyNunn I heard that Montreal recently paid off the 1976 stadium... as in in the last few years.
 
the olympics are a scam
 
user55340
(apparently, 2006 was when the $1.5B debt on the stadium was paid off)
 
> * There should be continuous access to sufficient food and drink of "high quality". Light snacks and canapés are not good enough
 
user15026
@MichaelT It was within the last few years, yes
 
user55340
2:27 PM
@AshleyNunn So... any city want to get saddled with another bigger debt for today's olympics?
 
@Ampt: The API design should drive your code, not the other way around. If you allow additional needed functionality to pollute your external API with new methods (because it needs to be tested properly), then your external API never stabilizes.
 
user15026
@MichaelT Yeah, I don't think so. It's gotten so ridiculous.
 
Microsoft does this all the time. They have a Split method on the String class, but it is driven by a private SplitInternal method. It's done this way to allow overloads like "discard duplicate delimiters" to modify the split behavior.
 
@MichaelT those ridiculous demands for private roads and free booze are really adding insult to injury
 
user55340
Look into the Java classes.. they're full of private methods... and some that don't exist between 1.6 and 1.7.
 
2:30 PM
The last olympics in Sochi was so publicly corrupt. Now anyone associated is going to be seen as being a part of it.
 
Ah, never mind. People are talking about the Olympics now.
 
user41796
What amazes me is the claim that the 2022 Olympics only need $5.4B but the hosting costs haven't been that cheap since before 1988.
 
You don't mean to say that the IOC is corrupt do you?!
 
user15026
@MattS. That and the demands at this point are so over the top - its like they are trying to drive countries into large debt piles.
 
user55340
@Ampt Just slightly less than FIFA.
 
user41796
2:31 PM
@MattS. apparently even the local mafia were killing each other off for a bigger slice of that pie
 
user55340
Hmm... wonder if Quatar can make a bid for the 2022 winter olympics.
 
user15026
@GlenH7 Yeah, that's about as likely as...a. very unlikely thing.
 
@RobertHarvey Getting into a religious war over testing methods wasn't on today's agenda. Sorry :)
 
user55340
You should put your private public internal methods in a #reigon so that you know not to touch them with refactoring.
 
user41796
@MichaelT apparently only non-democratic countries are willing to tolerate the associated and requisite graft that goes along with hosting
 
2:33 PM
Ocean's IOC
 
user41796
@Ampt - but who is getting fleeced? I suspect it's the host countries themselves...
 
Anyone who pays taxes would be a start
 
The people living in the host country.
 
user41796
Brazil demonstrated that FIFA and the World Cup isn't far from having the same sorts of issues.
 
user55340
@GlenH7 Just wait for Qatar. Its on track to make Brazil look great.
 
user41796
2:36 PM
@MichaelT The stadium design generated enough chatter, that's true.
 
user41796
Appalling work conditions are the norm there, and there's an utter contempt for immigrant labor. I am doubtful we'll see much news about the human cost involved in construction.
 
@GlenH7 I think that would depend on what kind of news services you frequent. I'm sure reddit will be aflutter with allegations, there'll probably be a few trending hashtags and maybe CNN will put out a story about the hashtags and gloss over the details of why they exist
 
user15026
@Ampt This is likely.
 
user15026
And tumblr will yell things at each other
 
Using gifs
 
user15026
2:40 PM
In their yelly vaccuum
 
user55340
@Ampt soccer vs football gifs.
 
@MichaelT "It's Called FUTBOL".gif
 
user41796
with cats?
 
plenty of those too
 
user41796
Just skimmed an overview of the bribes involved with the 2022 world cup.
 
2:41 PM
When are we getting an IOC replacement again?
we really ought to make it say (censored) instead of (removed)
maybe even [REDACTED]
 
user15026
I vote [REDACTED]
 
you're a mod so your vote counts for twice as much, and the rest of us regular's votes count for nothing.
 
user15026
So basically I win?
 
user55340
@AshleyNunn hard to tell, your vote is redacted so I can't see what it is.
 
Well since twice of nothing is nothing...
 
user55340
2:52 PM
That 'why make a function with no parameters' question is oddly rewarding.
 
user55340
54
A: What is the benefit of a function without parameters which only calls another function

MichaelTPlease pardon my memory if I have this incorrect... Javascript isn't my preferred implementation language. There are several reasons why one would want to have a no arg function wrap another function call. While the simple call to window.alert("Hello"); is something that you could imagine just ...

 
Why is it that every single applicant that sends us a code sample sends us 150+ line methods (often this method is main)
i beat myself up over 20 line methods
 
user41796
@durron597 You missed yesterday's conversation about that subject
 
user41796
The short answer is that they don't know any better
 
I guess the people that are looking for work aren't the people who write clean code
 
2:59 PM
Don't forsake them all
 
anyone interested in being a Java programmer in houston?
 
You hear about the recruiter who asked for LAMP devvelopers on github?
 
no
 
user41796
I would be more worried with how they are encapsulating their logic than the length of the function. But that's me
 
@GlenH7 if the entire business logic of the application is in main?
well i guess then it's not very encapsulated either :-P
 
user41796
3:02 PM
Is it less than 8000 lines long?
 
yes
 
user41796
ok, doesn't beat my horror story then. :-D
 
someone sent you an 8000 line code sample?
 
user41796
(waits for that to sink in)
 
user41796
@durron597 oh hell no. That was production code.
 
user41796
3:03 PM
I forget how many functions I refactored that in to
 
user41796
Apologies to anyone that spat out coffee after reading that.
 
the winner is still the guy from the stackoverflow thread
that would do things in C# like this.x = this.x
and if you removed the line it would break
because so much other crap happened in the property setter
 
user41796
Yeah, can't argue there. My IRL stories are comparatively tame in comparison to some of the worst
 
> There's an if statement that checks the StringSplitOptions enumeration, and runs one of two different methods depending on which StringSplitOptions setting is specified. The two methods are called InternalSplitKeepEmptyEntries and InternalSplitOmitEmptyEntries. They are pretty much two completely different implementations.
 
System.err.println( "Theres a gremlin in the server: "+ref.getClass() );`
 
3:08 PM
So yeah, I suppose you can test those from the public API by throwing the switch on the enumeration.
 
that is simple enough to just assume it's a black box
 
I'm not so sure that would always be the case.
One day in the future, when everything becomes solid TDD, someone will come along and say "y'now, it would be nice if we could test our internal methods independently of the public API." And a new best practice will be born.
 
user41796
@RobertHarvey Didn't know about the Telerik product, thanks for mentioning that. I've been having a horrible time figuring out what the RabbitMQ .NET client is (or isn't) doing. I'm going to give that a try.
 
Can I just say that a lot of times respected open source projects aren't any better
 
Better than what?
 
user41796
3:15 PM
 
Mmm that was not the best way to phrase what I meant
Many respected open source projects don't encapsulate themselves well and make it difficult to modify, extend, etc.
Specifically, I was trying to figure out a way to easily make a JUnit runner that would allow one to easily use PowerMock and Jukito's features at the same time
but chaining runners is just... not easy.
 
I wonder sometimes if mocking isn't overrated. There are things in ASP.NET MVC that are difficult, if not impossible to mock, and one of ASP.NET's key features is supposed to be testability. Seems to me like it would be better to write code in a way that you could test it without having to write mocks at all.
 
user41796
There you go being all heretical again...
 
Well, the only reason this even came up is that I have a third party class "MyGodObject"
and I wrote a class ConfiguredGodObjectFactory
And the test cases tests that the MyGodObject's methods were called the way they are supposed to be called
but the problem is that MyGodObject is full of final methods so they can't be mocked without PowerMock.
And I'm not really sure how I would test such behavior without using a mock. But if you have ideas, I'm interested.
 
[looks suspiciously at that class name]
 
user55340
3:24 PM
@RobertHarvey You're right... should be OhMyGodObject
 
I probably should have wrapped the MyGodObject in some sort of adapter but this was some of the first code I wrote two years ago and I haven't refactored it yet. Seemed more trouble than it's worth. But maybe not.
 
user41796
Apropos of nothing, why doesn't SE show when someone is Q or A banned? ie. why is that information restricted to mods+ only?
 
Because if they always showed it, people would figure out how to game it.
 
user41796
ah, ok. That makes sense
 
user41796
It's an anti-gaming measure as opposed to a protecting personal information measure.
 
3:36 PM
They show quite prominently when someone is suspended, though they are always vague on the details.
 
user41796
The rabbitmq .net client is making me feel stupid. (Not hard, I realize) I can't figure out how any code is being executed when a particular function is being called.
 
magic
 
user41796
@RobertHarvey Not explaining the reason why in that case is understandable
 
user41796
@ratchetfreak certainly seems that way.
 
Hmm, are they using Reflection anywhere?
 
3:37 PM
it's what the m stands for ;)
 
user41796
@RobertHarvey Honestly, I don't know / can't tell. grepping the entire source tree doesn't give any hints with this particular function
 
Let me guess... No code comments either?
 
user41796
Nope. Trying to find a direct link into their repository to show
 
@RobertHarvey It's self-documenting
 
And apparently self-running, too.
 
3:40 PM
@GlenH7 you could grep for the reflection API?
 
user41796
 
user41796
@ratchetfreak sure, what is it's name again? I don't recall offhand
 
user41796
Yep, they've got System.Reflection referenced in a number of places.
 
user41796
Sorry I linked the interface, not the implementation. hg.rabbitmq.com/rabbitmq-dotnet-client/file/c7a72d4a6f20/…
 
user55340
Btw, anyone give any thoughts on what could be done with advertisements to meta posts?
 
user55340
3:52 PM
@RobertHarvey I wonder if you could put an advert to Eric Lippert's "how to debug" posts on SO...
 
4:03 PM
@MichaelT It's already cited in stackoverflow.com/help/how-to-ask.
(granted, it is a bit buried there)
 
user55340
@RobertHarvey But no one ever reads help... to have a nice, colorful thing on the side that says "learn how to debug by Eric Lippert of C# fame"...
 
user55340
4:23 PM
Today's time waster... a wind map. Note, it looks like that 'kitchener' area is trying to steal all our air.
 
user15026
@MichaelT It's miiiiine you can't have it
 
user55340
@AshleyNunn I was expecting "and its from Chicago and Columbus?! So that's what I'm smelling..."
 
user15026
@MichaelT I am in the office, the air gets recycled! I can't smell it :P
 
user55340
@AshleyNunn just your co-workers exgassings...
 
user55340
@RobertHarvey The other thing one could do is to make an advert for /review/close
 
user55340
4:32 PM
See if you could get SE to actually make that a proper campaign that shows up for users with > 3k rep.
 
5:06 PM
Hmmm
I pissed off the mainframe COBOL devs at the bank
 
@maple_shaft I would think that anyone who is still a COBOL dev would just be pissed off all the time anyways
 
I made a comment to the effect of, "We are trying to avoid mainframe development in the future because it is fundamentally at odds with the enterprise architecture goal of continuous integration"
They basically got pissy and said, "You frankly need to stop talking about things you don't understand!"
am i wrong?
or are they just scared?
 
user41796
@maple_shaft a little of both
 
CI is a thing in Mainframe dev?
I guess I dont really know
 
user41796
Many financial institutions pay through the nose for COBOL programmers because they have a huge codebase that can't affordably be replaced.
 
5:10 PM
Maybe I am coming across as a know it all
 
user41796
So you pushed some buttons when you said "we're trying to cut back on your development efforts."
 
user41796
But ignoring the fact that you torqued them, there's nothing that says COBOL development can't be done in a CI manner
 
user41796
And to be honest, that's the type of app that could really benefit from some of the benefits of CI and an automated test suite.
 
user41796
In a nutshell, CI basically says "we're gonna build every time we change the code base." The point being that you want to make sure the build didn't break and all the tests continue to pass.
 
user41796
5:13 PM
There's nothing within that where it says "but we're ignoring legacy languages"
 
user55340
Or the "when you do want to do a build, its ready rather than taking hours to get to that point" or "when we have a build, we can easily and quickly deploy it to some environment"
 
user41796
You may not currently have a very robust test suite for that app / those apps, but you can still run the compiles and make sure it builds
 
user41796
@MichaelT Yes, most definitely
 
I am probably coming across as an asshole then
 
user41796
CI is a process that encourages smaller commits as they are less likely to break stuff
 
5:14 PM
I just assumed that automatic deployment was a key part of CI
 
user55340
At Employer^^, had a jenkins build (system had no tests) that would allow us to push to dev from a consistent build environment so we could test... and the 'allow us to push to dev' was a single button click.
 
@psr These mainframe guys are giving me a hard time. I lost the room
 
user55340
Automatic deployment isn't always necessary or part of CI. Automatic builds is. Those facilitate other parts that can make other things much easier and practical.
 
@MichaelT What about builds though? Source code analysis in Jenkins is one thing, but what about verifiying compilation?
What about COBOL unit testing?
What about mocking?
 
user55340
 
5:18 PM
I am still learning COBOL but I don't see where that is natural or easy
not yet
 
user41796
@maple_shaft COBOL predates a lot of DI concepts. I wouldn't hang your hat too much on ever seeing that.
 
user55340
It may require some architecture adjustments... and those might not be practical... but it exists.
 
user55340
But automated testing is one of those things that CI makes easy, but doesn't make CI.
 
user41796
@MichaelT Alternatively you can adulterate the unit test approach or principles in order to get the code exercised.
 
user55340
You might make better inroads into it with static analysis of cobol (which also exists) as part of the build process.
 
user41796
5:20 PM
For example, you may say "meh" to only testing one thing at a time. You may have to test multiples and then provide a better error message that said "these were the parms and this is how it died."
 
@MichaelT The thing is that if I want this to happen then I would have to teach them how to do it. They wouldn't do this or change this
 
user55340
 
user41796
@maple_shaft because they don't see the value in it
 
user55340
@maple_shaft mainframes are very entrenched places.
 
user41796
show the value and you'll change their minds
 
5:21 PM
then management would say, why are they wasting their time with COBOL unit testing framework when they could be helping you moving this to middleware
 
user55340
Point out "it lets you be sure I'm not making stupid mistakes as I check in code - fewer things for you to watch over the newbie for"
 
@MichaelT Newbie? Cobol?
They are a bunch of greybeards
 
user55340
@maple_shaft you're learning cobol...
 
touche..
:)
 
user55340
Let the graybeards know that by implementing these tools they don't have to look after you as much.
 
user41796
5:22 PM
Another advantage of unit testing in COBOL is you're effectively creating specifications for the code
 
user41796
So when you go to port that code to another language, you have specifications you can use to validate it's operating as required.
 
Hmmm... I was told to learn it to help rewrite modules where I can, not so that I can write modules
 
user55340
Also point out that at some point they will retire, and management will need to either spend $$$$$ to replace everything with Java (or C# or whatever) or hire new programmers and teach them cobol... without the mentorship of the graybeards.
 
user41796
Turning that around, without good unit tests or a specification trying to port a large, legacy application is a huge recipe for disaster
 
I am just trying to learn all the hotkeys in this "terminal-ish" ancient feeling interface
 
user41796
5:23 PM
@maple_shaft How can you safely rewrite without validating what it does?
 
@GlenH7 I can't
 
user41796
@maple_shaft That's just being ageist... :-D
 
user55340
Thus, by putting these in place it allows you to (a) concentrate on the important parts of mentorship that the static analysis and testing won't cover and (b) make sure that the hard bits of the cobol code won't break as easily when they retire.
 
I am in over my head
 
user41796
Welcome to enterprise development
 
user41796
5:26 PM
You're not working at it unless you're in over your head.
 
Some of the mainframe guys have expressed extreme interest in learning Java from me though
 
user41796
And I can verify the veracity of that statement from years of enterprise dev and having been fortunate to work with honest and open devs at all levels.
 
They are so out of date I am not sure where to start
 
user41796
@maple_shaft Boom! You win; they win.
 
OOP? JVM concept?
 
user41796
5:27 PM
COBOL isn't as bad as you think. The whole concept of Playbooks was an early adaptation of DRY before DRY existed.
 
user41796
I can't remember if COBOL has structs or not.
 
Whats the difference between Playbook and Copybook? They say the word Copybook a lot
 
user55340
Employer^^ is looking at switching their back office programs from Progress to Java because the difficulty hiring/training Progress developers (much easier to get kids out of the local university that know Java rather than try to persuade them to code Progress)
 
user41796
@maple_shaft I'm guessing equivalent.
 
user55340
Side bit - there's also a Cobol for JVM if you want to play that way...
 
NEAT
 
user55340
Another article touching on it: jaxenter.com/…
 
user41796
COBOL tried to commit seppuku but too many existing applications cried out for its continuance.
 
user55340
So if corporate does want to move to Java some day, its possible to move some of the hairy parts of the cobol as cobol rather than recoding them.
 
user41796
And many times the risk of the port going wrong simply outweighs any other consideration.
 
user55340
5:32 PM
(especially for banks). Though do realize those pieces of big iron they've got likely aren't going away... their IO is just too good compared to even the cloud things of today.
 
user55340
I worked with HSBC (now Capital One) in credit card applications and had to make sure I put a flag in a message that said "this is ascii, not ecbdic. And when you start thinking of the iron handling all the credit card processing for a large bank like that... their old iron preforms just fine.
 
The current plan is to seperate out business logic from all cobol modules around payments through wire and ACH systems, leave modules that take flat files and interface with other mainframe systems, and hide these interfaces behind an ESB that fronts web services and MQ in a consistent format
 
user55340
@maple_shaft I'm going to guess IBM WebSphere for the MQ?
 
WebSphere MQ yes
 
user55340
That's what HSBC had for their backend too.
 
5:34 PM
the part that worries me is that batch processing will be happening in middleware now
 
user41796
@MichaelT Pretty much guaranteed to be websphere mq if a mainframe is involved.
 
millions per day
 
user41796
@maple_shaft running on what platform?
 
user55340
The nice thing about the MQ is that they can scale it if they want to.
 
WebSphere
So the vendor software part is horizontally scalable infinitely
it uses ExtremeScale software for object caching
There is going to be a LOT of servers I think
the vendor softwae is a JavaEE app that scales well
 
user41796
5:38 PM
@maple_shaft Mebbe. Don't underestimate the transactional processing power of the mainframe though
 
I don't believe me...
that is what worries me
Some big institutions successfully use this in place of mainframes
but we are going to be doing some batch processing further in middleware that used to happen in COBOL on the mainframe
 
user41796
but that said, you can build an equivalent system on an open platform stack. Just takes more fiddling with it. There's a high frequency trading firm over in London that uses Java for their base language along with some queuing software. I don't have the link, but may be worth trying to find.
 
user41796
@maple_shaft Find the stuff that's stupidly parallel and make sure you run it in parallel.
 
I know its possible, but I am worried about how many servers we will need in production to deal with this load
may need 100 servers
or more
 
user41796
@maple_shaft what chipset?
 
user55340
5:42 PM
This is a bank... they'll pay any amount of money to continue to handle money that they make money on.
 
user41796
x64?
 
6 million ACH transactions and wire transfers a day
thats some serious thoroughput
 
user55340
Read the VAX War and chuckle...
 
user41796
shouldn't be an issue on a decent couple of iSeries or AIX boxen
 
x64 RHEL probably 16 core ? RAM
 
user55340
5:43 PM
> VAXen, my children, just don't belong some places. In my business, I am frequently called by small sites and startups having VAX problems. So when a friend of mine in an Extremely Large Financial Institution (ELFI) called me one day to ask for help, I was intrigued because this outfit is a really major VAX user - they have several large herds of VAXen - and plenty of sharp VAXherds to take care of them.
 
user41796
@maple_shaft beefier unix servers are out of the question?
 
virtualized but on dedicated hardware
No AIX is not out of the question
 
user41796
I'd look into it simply to reduce system administrative overhead
 
user41796
I'm not saying you can't run that many transactions on x64 based hardware. But there are more powerful tools out there for not that much higher on the price curve
 
user41796
and when you factor in labor overhead costs, different chipsets become more attractive
 
5:45 PM
@GlenH7 Any ideas on some serious fucking chipsets for serious data crunchers?
 
user41796
IBM's Power chips are fixed, capital costs. Labor is forever.
 
user41796
@maple_shaft (Not joking) how's the capacity on your mainframes and what's their vintage?
 
Don't know. I can't tell you the first thing about them
 
user41796
I also can't speak highly enough of iSeries for transactional processing without being in the stratosphere of mainframes
 
I know that the wire system though hasn't been upgraded hardware wise in over 15 years
 
user41796
5:46 PM
but iSeries is a different beast of an OS and I'd hesitate to recommend that without inhouse knowledge
 
@JimmyHoffa is a virtual treasure trove of transaction mangement, right Jimmy?
 
user41796
So I'd look at the servers backing AIX
 
I am not allowed to make decisions like that. The architects and infrastructure folks choose this for us
based on our needs
I can certainly suggest though
 
user41796
I would want to know current utilization and age of the existing mainframes. You can do a lot of amazing things even in modern languages on them.
 
user41796
The big selling point there is it's already a sunk cost and you already know how to administer the machine
 
5:49 PM
@GlenH7 I don't disagree with you, but I would suggest possibly virtualizing other OS's with modern software on them which I don't believe is going on today
 
user41796
But I'd be willing to bet that IBM's Power systems running AIX would also hit the sweet spot if you're looking to acquire hardware too. I'm assuming you have in-house knowledge of unix admin
 
user41796
@maple_shaft pretty darn certain that won't be an issue on the mainframe
 
they have some archaic text based interface as the means to use it
 
user41796
@maple_shaft because that's their choice. There are other options. :-)
 
they don't even allow X11 or graphical programs on AIX/RHEL
 
user41796
5:51 PM
How long did it take to get the JVM installed on your desktop? Think about if the existing developers would want to fight for new software on theirs....
 
@GlenH7 Still waiting :)
 
user41796
So how long would it take for approvals to open new ports on the firewall and get new software installed?
 
I used "dark arts" to get it installed in my Users directory though without admin rights
 
user41796
I'm just saying that those options aren't in place possibly because the developers didn't feel it was worth fighting for
 
my company lets me have enough admin rights I can install it normally, then copy it to my users directory heheheh (for those I can't install as user only in the first place)
 
5:52 PM
@GlenH7 If I put the request in... oh an hour from now... then probably sometime in the next 20-25 years
Ya know...
Innovation
 
user41796
So there you go... :-)
 
user41796
That said, working remote over a terminal connection tends to be rock solid.
 
user41796
Been there, done that.
 
They had a town hall with the CIO for the Vision Innovation initiative program they keep talking about
I asked a question in the forum, "Seeing as how developers and QA are completely unable to do their job without admin rights, a 45 day SLA for approving new desktop software, and no development sandbox we can use as an alternative, how soon until we have a solution that will allow us to do our jobs and be innovative?"
 
user41796
@maple_shaft If your place is anything like the institutions I've known, you may want to work on saying that in more subtle ways.
 
user41796
5:56 PM
If only to keep your sanity
 
The CIO went into some rant about the private cloud and how QA can spin up 10 test machines to do performance testing with all of the software they need already prebundled on it...
 
user41796
@maple_shaft So he's advocating for stealth IT?!
 
Then said, "Yeah I understand that being a developer here can be a little kludgy................... Next question!"
 
user41796
@maple_shaft About par with what I'd expect from an exec.
 
A.K.A "I don't give a fuck about developers. Fuck you all! You are an enormous security risk on our network you liability ridden fucks!!!!"
Innovation
 
user41796
5:58 PM
@maple_shaft Or they are trying to be understanding in that developers are human too and that they are prone to mistakes as part of the development process.
 
user41796
Although with that in mind, I'd put 'em all off on their own private vlan and isolate them from the rest of the network. Then they can have more permissions and are less likely to frak things up.
 
@GlenH7 Maybe if they didn't hire so many worthless H1B's that don't know what they are doing then it wouldn't be a problem would it?
 
that sounds miserable
 
user41796
Sun Tzu's Art of War may be a good read for you
 
user41796
@maple_shaft nah, that's not even it.
 
00:00 - 18:0018:00 - 00:00

« first day (1491 days earlier)      last day (3502 days later) »