• AP (unregistered)

    Not sure about passing messages but frist one to comment!

  • ray10k (unregistered)

    Enough red tape to strangle an elephant with, a business that's more concerned with following it's self-imposed rules over working efficiently... Any chance this was some Government job, or was it just some company that thought it was on That Level?

  • bvs23bkv33 (unregistered)

    that is good like you have half year of initial analysis, analysis, business specification, technical specification, functional specification, unit test planning two days you write two pages of code then you have half year source code management, test environment installation, deployment to test environment, unit test performing, impact analysis, explaining to testers that they went out of case coverage and already testing functionality nobody talked about, security approval, freeze period and so on

  • El Dorko (unregistered)

    Sounds awfully lot like something we might do here... We have a fairly small company that has almost as many servers as users, and we only use cloud-hosted software (office365 and the likes). But everything has to go through channels of "security" and validations and dmz's and whatnot, and we more or less have 4 to 20 servers per each externally hosted system, just to route the traffic "securely" through, or something. And yes, every single piece of hosted apps is accessed through HTTPS so meh, but yay, "security"! Why be smart when you can be corporate-y?

  • Mike (unregistered)

    I'm getting flashbacks to a previous job at a Large Financial Institution...

    Verrrrry similar nonsense to just set up a Subversion server...

  • Aspie (unregistered)

    This sounds like IBM trying to be efficient.

  • Flag1 (unregistered)

    I'm beginning to suspect snoofle is one of the horsemen of the Apocalypse: Death, Famine, War, Conquest and DevOps. Wherever he appears, doom is nigh. If he ever shows up at your workplace, run!

  • (nodebb)

    I worked for an ISDN videoconferencing company a few years back, shortly after TBL first rolled out www.

    I stumbled on an IP videoconf proof-of-concept system. My boss said, "call a meeting and do a demo." As you might guess, the company had a vital interest in understanding this sort of thing. Short term: figure out how to sell against it. Longer term: figure out how to do it better.

    But, the network administrators would not allow the proof-of-concept system to pass through the company firewall. So, no demo at the office. A couple of people came over to my place to see it. At home I was an early broadband adopter. (LAN City modems, anyone?)

    That company was acquired by a competitor three years later. I learned how to explain what I need to executives, and get them to instruct the network administrators to cooperate.

    Sometimes the best way to communicate across a firewall is by using a couple of trebuchets to fling CD-ROMs over the firewall.

  • ichbinkeinroboter (unregistered)

    wow. I did a big project with DTCC, but I forget where and I forget what. (I have been a contractor for too many banks.... ) I just rememember the huge manual... Thankfully, I don't recall any hideous bureaucracy.

  • Ano (unregistered) in reply to Flag1

    Thank God he's retired then. Although we could really use someone like that where I work... Would be the end of the world as we know it, so Apocalyptic Horseman would even be the correct title.

  • Noob (unregistered)

    Pretty common in enterprises. No deprtmemnt wants to be the one that fucks up, so they are stringent in terms of reqiurements.

    That said its nbot hard to fix. Get one person (no more) from each department that is senior enough to sign off work, get them around a table, agree its a test, and then spec the work.

  • Jim B (unregistered)

    Is this the bank I used to work at?
    A BA there once calculated that it would take a minimum of 3½ months to move a two-line code change into production. It required (and this was for a 'simple' change) - business requirements, technical requirements, business-signoff, coding (was already done but still needed a line on the 'project plan'), functional testing, regression testing, performance testing, build engineers (in India) that were the only ones to move between the 5 (five!?) environments before production (they could only do after 10:00 PM and only certain environments per night). All this for something like changing the color of a button for a certain condition (hence the 2 line change). Where I'm at now we can compress this to one day. I only need to get someone to move code to production as we can do the rest in the lower environments. Much saner.

  • Kashim (unregistered)

    So basically just Waterfall model + design-by-committee + everyone covering their butts. At some point, I just go to someone who cares about financials, and show them the cost of what we've done, how much money I can save them by doing things the faster. If I can't find someone who cares, I write a well founded article to the CEO. In the end, it's his money, and if he wants to spend millions on bureaucracy, then that's cool, but if I can show him a way to save all that money, and how much it will cost for us to test it, then maybe I can get a raise, and everyone else can get a talking-to.

  • snoofle (unregistered)

    This place was the bureaucracy to end all bureaucracies - worse than every place I've ever worked - combined.

    There was a second level manager who was being forced to try this new-fangled messaging thingy, but he didn't want to do it, so he used the bureaucracy to kill it. I actually almost got around it on sheer persistence. When he saw that it was about to bear fruit, he changed the requirements forcing me to back up about 9 months, after which I very publicly called him out on it, and then left.

  • FormalWare (unregistered) in reply to Aspie

    Like this attempt, you mean?

    http://www.cbc.ca/news/canada/ottawa/ibm-36-5-millions-dollars-more-phoenix-system-1.4643809

  • desertrat (unregistered) in reply to El Dorko

    Potemkin Security, I call it, looks very good and impressive to PHB's, pretty much useless for it's task.

    Like some id10t's where I work (in the Infosec office, no less) who habitually post emails with x's for the 'tt' in http and https. Ooh clever, except...some browsers skip right over it, and end users just think that these infosec people are bloody idiots they can't even properly insert a hyperlink!

    When I point out that a) bad actors will never do this and b) it focuses the end users attention on the wrong part of the address, they look at me as though I've sprouted three heads and am speaking in authentic frontier gibberish. And go on doing it...

  • Sole Purpose of Visit (unregistered) in reply to Kashim

    No, basically nothing like the Waterfall model at all.

    Which doesn't even exist.

    But thank you for playing.

  • (nodebb) in reply to Aspie

    This sounds like IBM trying to be efficient.

    Funny, I was thinking that. Also funny is, their implementation of MQ is just about as clunky and cumbersome, even if it's all software driven.

  • siciac (unregistered) in reply to snoofle

    It's a vicious cycle as that kind of shit drives out the people who don't want to deal with the bureaucracy and can find work elsewhere.

  • J (unregistered) in reply to OllieJones

    This has made my day. I have to write this down somewhere.

  • faded (unregistered)

    Is there a chance the company with the recto cranial sublexation was General Electric? I did project with them years ago. The names of the departments and groups in the original post sound awfully familiar to me.

    Their corporate IT folks actually wanted to inspect the window locks in my office to see if they were secure. When I heard that I went back to the customer and told them what was happening. Then I doubled the amounts on the contract. The customer did not blink. The agreed to pay because they knew that corporate IT was a pain in the ass.

  • snoofle (unregistered) in reply to faded

    Definitely not GE. It is more of a place that deals with teaching.

  • (nodebb)

    At some point, SOMEBODY should have said "This is totally impractical and expensive, so we ought to go back to doing it the way you WERE doing it - even though that's the wrong way to do it."

  • Jim Tonic (unregistered)

    Sounds to me that there wasn't really a problem to be fixed to begin with. Sure file-drops and cron-jobs don't sound as fancy as MQ, so the managers can't show off. But if it ain't broken, don't fix it.

  • Bappo (unregistered)

    Do we work in the same place ?

  • (nodebb) in reply to Jim Tonic

    But if it ain't broken, don't fix it.

    File-drops and cronjobs are just about as broken as "it doesn't work". Absolutely NO reason to have that running with today's tech.

  • DrPepper (unregistered)

    Just got done working on a project to give people access to "FAQs". Each department in the company could contribute FAQs and answers, which would be stored in a db and queried by the end users. Simple, off-the-shelf type stuff. By the time I left the project, there were three separate groups meeting (I was part of each of the groups); each group involved several teams, with 2-3 people for each team -- probably 30 people all together. Intellectual Property people, network security, developers, designers, content writers, department heads, ... for a project that should have taken maybe a day or two to put together.

  • MG (unregistered)

    This sounds an awful lot like a company we'll just call "Black and Yellow Security Software," where getting a firewall rule implemented often took months.

  • (nodebb) in reply to Bananafish

    But if it actually does work (let's assume the number/size of files is small enough that it does), then what justifies the cost of replacing it with a MQ? (Let's also assume that the MQ isn't drowned in WTF as in the original article.) Sincere question; I know basically nothing about MQ except skimming the Wikipedia article. I assume the main point is bandwidth management and/or centralized overview of multiple input/output sources.

  • Zenith (unregistered) in reply to emurphy

    I never set up an MQ installation but I did configure a few additional queues and write interop code against its DLLs. The benefit of MQ over file drops was that you could rerun old messages and not have to clean up after it as much. That said, it's really easy to write a service that does exactly what MQ does, or at least the subset that you need. But then you run into the dummies that use a worthless support contract as a safety blanket. If something I wrote goes down, I can fix it in minutes. Getting support from a vendor (thick accents and all) can take hours or even days.

  • eric bloedow (unregistered)

    somehow this made me think of an old story where programmers were forced-or thought it was a good idea-to send information from one computer to another by EDITING THE WINDOWS SYSTEM REGISTRY! absurdly inefficient, and downright dangerous.

  • medievalist (unregistered) in reply to Jim Tonic

    File drops are a great and very future-proof way to pass data between institutions. If you use passwordless SCP with very large keys it's the current state of the art.

    Cron jobs, however, are only a good idea for scheduled activity, polling is a horrible way to implement an event-triggered process.

    If you want to instantly react to the arrival of data, instead of cron use iNotify on linux (gidget for example, http://www.typinganimal.net/code/gidget/) or if that's too high-tech and scary just monitor the file transfer logs (SEC for example, https://simple-evcorr.github.io/).

    File drops work just fine - and are trivially easy to get through both firewalls and bureaucracies.

  • death the kid (unregistered)

    Nice, very Nice. They're bureaucrats i don't trust their words..

    We can always prove mathematically that they're dumb, i mean very dumb....

Leave a comment on “Passing Messages”

Log In or post as a guest

Replying to comment #:

Log Out

« Return to Article