• TheCPUWizard (unregistered)

    Just a note... "Software Engineering" is distinct from "SDLC" is distinct from "ALM"... Each can be abused (i.e. use the name, but not the "real" parts), but also they can be utilized together.

    What is interesting [to me] is that most "App Dev" eschews SE because of the rigor, then is surprised when they have "surprises" because they tried to gloss over them...

  • Steve_The_Cynic (nodebb)

    Of course Software Engineering is distinct from Synchronous Data Link Control.

  • Sceptic (unregistered)

    The best software creation process called by any jingle word whatsoever won't amount to much when, at the end of the line, badly trained, underpaid, and myopic “debuggers” peruse tiny snippets of code one at a time and sprinkle myriads of “patches” throughout without even attempting to see the big picture. Thus, the software creation process might as well start off in the same bumbling, sloppy, short-sighted way as it will end, and since everyone knows this at heart – it does.

    For decades now, all software creation has been a sophisticated muddling, with no promising solution in sight.

  • isthisunique (unregistered)

    I once setup one of the variants falling under HP ALM's softwate suites to try it and all I can say is stay away. Install documentation is wrong, scripts look like they're written by monkies, bloat is extreme and it's extremely hard to even get the evaluation or free version working without installing tons of other bloated commercial software.

    Installing it on a Linux machine with all the requirements such as Oracle DB (express version) pretty much destroys the machine with kak all over the place of installers that don't say what they're doing nor are uninstallable, just reinstall Linux over it.

    I wouldn't normally have bothered but someone else had experience using it and I wanted to see what it was about.

  • my name is missing (unregistered)

    My sister once worked for a company that sold a tool to assist programming teams in shipping products faster. At a meeting she asked when a particular release would go out and they gave a date far in the future. She then asked if they used their own tool and they admitted they did not.

  • Brian Boorman (google) in reply to Steve_The_Cynic

    If you're going to talk SDLC, then done forget about it's partner - HDLC.

  • Steve_The_Cynic (nodebb) in reply to Brian Boorman

    Indeed. I hadn't forgotten it, but it didn't seem relevant to the joke.

  • comments.Count.ToString + "th" (unregistered)

    If you're gonna play lipservice to anonymization, at least remove Mercury from the error message...

  • comments.Count.ToString + "th" (unregistered) in reply to comments.Count.ToString + "th"

    Also, isn't it nice that they still haven't purged their own code/error messages from references to Mercury, and the merger was in 2006.

  • Chronomium (unregistered)

    DECLARE @sql AS VARCHAR(4000)

    This could be a standalone article in the Representative Line series.

  • Bananafish (nodebb)

    This shouldn't have gone out, but it's probably an edge case in a code generating loop. I have seen this before where an interim dictionary was compiled but no data changes were applied.

  • isthisunique (unregistered)

    This shouldn't have gone out, but it's probably an edge case in a code generating loop

    I doubt it. When I set up ALM it was fubar. It was that moment I knew for sure everything I've believed about enterprise software is true. The install process was messed up with a lot of things out of order, that had to be rerun and SQL scripts you need to use not mentioned anywhere in the install notes.

    It's just not properly engineered and a legacy behemoth. I don't think they care about doing things properly at all because they just want support contracts for that. I've seen similar things with MySQL turning enterprise with rubbish like password validation as a default plugin and GTID replication being messed up where you need to create complex scripts with horrific hacks to do basic things (essentials in many cases) that should be simple.

    Setting up Oracle express wasn't any fun either.

    This is what happens in enterprise. The less streamlined a technical process is the more support contracts are required to get things done.

  • Rob (unregistered)

    I use this horrible tool for work.

    Let me just say, shame on them for releasing this product. I can't even list the amount of bugs. As a developer, I would be ashamed to put my name on this product.

  • Rob (unregistered)

    I use this horrible tool for work.

    Let me just say, shame on them for releasing this product. I can't even list the amount of bugs. As a developer, I would be ashamed to put my name on this product.

  • Oliver Jones (google)

    Well, it looks to me like our young friend "Bobby Droptables" grew up and got a job in our industry.

  • slavdude (nodebb) in reply to isthisunique

    Install documentation is wrong, scripts look like they're written by monkies, bloat is extreme and it's extremely hard to even get the evaluation or free version working without installing tons of other bloated commercial software.

    So... open-source, then?

  • Fnord (unregistered)

    Not only is the install documentation wrong, chances are that your Harry Peckherd contact will give you a previous version's instructions anyway.

  • siciac (unregistered) in reply to Bananafish

    but it's probably an edge case in a code generating loop.

    An "edge case" refers to extreme values that cause anomalies in floating point when evaluating otherwise correct mathematical formulas. For example, Heron's formula for area is perfectly valid mathematically, but with floating-point and needle-like triangles it gives incorrect answers.

    When people are just hacking at code until it satisfies a few unit tests, the perfectly ordinary inputs that fail are not "edge cases."

  • My Name Here (unregistered)

    As far as the script goes, the whole bit about dropping the index is nonsense: the where clause matches no rows, so @sql will be set to null. The test against '' (which isn't showing because presumably the less-than-greater-than version of not-equals was used) will therefore fail so @sql won't be executed.

  • Dave (unregistered)

    They have a long history of this. Excerpt from a comment from some code I maintain:

    "Kudos to Harry Peckherd's QA department for managing to ship a select() that breaks its own prototype".

  • MiserableOldGit (unregistered)

    Having had to use ALM (as in this HP/Mercury abomination) at the last gig, I can confidently say that if a patch managed to trash the fucking thing so it couldn't be used that would be the best feature upgrade they'd ever issued for it.

    When I enquired why we bothered with it when something else, anything else, even nothing at all would be better, the only coherent answer that came back boiled down to "because we spent so much buying it".

  • jerepp (unregistered) in reply to MiserableOldGit

    Anything... even your classic folder with .Old, .newOld, .DontUse, .Better, .newBetterer, .monkeyfarts

  • airdrik (unregistered) in reply to My Name Here

    Well, if you are going to trim that, you might as well truncate the script after the first line since everything else will fail or do nothing with the table having been dropped (and they are likely unnecessary anyway as those constraints would be dropped when dropping the table; and if not then the dropping of those constraints needs to be executed before the drop table).

    the @sql part at least allows the script to work even if the objects involved don't exist (though better would be to wrap the dropping of the index in an if block rather than doing dynamic sql).

  • MiserableOldGit (unregistered) in reply to jerepp

    Ah, I see you inherited one of my earlier codebases ...

Leave a comment on “ALM Tools Could Fix This”

Log In or post as a guest

Replying to comment #:

« Return to Article