• Anonymous (disco)

    Never. We’re not buying this. Thank you for your time, Spencer.

    I wonder if people in the real world speak like that...

  • accalia (disco) in reply to Anonymous

    sounds like something i would have said. ;-)

  • DocMonster (disco)

    To be fair though that's how a lot of consulting/software firms work. You don't need to be an expert, just smarter than the idiots you're selling to so you can hype your services.

    At least that's how my job works. We sell consulting for an app that we only barely know how to use ourselves. It's almost literally "Are you too lazy to do your job? Let us do it for you!"

  • Steve_The_Cynic (disco)

    When they said "Memory Database" I had visions of new and/or modified data being held for unreasonable amounts of time (say more than a second or so, but more likely all day) in server memory so it would be quick to access. And so it would go up in a puff of vaporware if the server process crashed...

  • Hanzo (disco) in reply to DocMonster
    DocMonster:
    To be fair though that's how a lot of consulting/software firms work. You don't need to be an expert, just smarter than the idiots you're selling to so you can hype your services.
    Oh, is that what they mean when they say they need an architect. They need a marketing person that they on an engineer's pay.
  • RaceProUK (disco)

    I like how two users managed to comment on this two days before the article was published :confused:

    And if a random outsider can speed up your queries by a factor of 119, then, yeah, I don't want that software. Not even if you paid me $5,500 to do so.

  • John_Imrie (disco)

    TRWTF is that Management still by this kind of crap.

  • accalia (disco) in reply to RaceProUK
    RaceProUK:
    I like how two users managed to comment on this two days before the article was published

    mark the Articles category as watched in your preferences and you'll get notifications when paula creates these articles, they're often a couple of days before go live. ;-)

  • JBert (disco) in reply to accalia

    INB4 new Frist! wars

  • Anonymous (disco) in reply to JBert

    You probably are new here

  • CoyneTheDup (disco)

    We are sending Spencer.

    Spencer is an enthusiastic advocate of new technologies and presents an excellent sales presence. Not only can he encourage customers to adopt the latest technological solutions, but his effective leadership skills guide the customers on how best to adopt advanced solutions. His ability to emphasize the need for partnership is beyond parallel. You can be assured the greatest possible return on your investment as you develop your business to meet today's challenges.

    ("Bingo!")

  • DocMonster (disco) in reply to Hanzo
    Hanzo:
    Oh, is that what they mean when they say they need an architect. They need a marketing person that they on an engineer's pay.

    I've seen it far too often that the "consultants" just know a bit more than the clients, and aren't really experts on anything. So you sell services that you barely know because the person you're selling to doesn't know (or doesn't want to learn) how to do it.

  • Shoreline (disco)
    Our current solution works well enough...

    The value of the upgrade is never mentioned. I assume it could just generally be better.

  • abarker (disco)

    Apparently no one has caught this yet:

    vendor ÜberWarehouse to upgrade

    ­

    ÜberWarehouse wasn’t about

    ­

    something ÜberWarehouse called

    ­

    answers from ÛberWarehouse in person,

    ­

    upgrading your ÜberWarehouse solution.

    So 4 out of 5 instances of ÜberWarehouse use Ü, and the other instance uses Û. I demand a rewrite! Either more consistency or more discoursistency, but pick one!

  • FullPointerException (disco) in reply to abarker

    My pre-coffee vision isn't good enough to tell the difference. I had to read your quoted sections twice before I spotted what you were talking about.

  • chreng (disco) in reply to abarker

    Is that you, Rainman?

  • abarker (disco) in reply to chreng
    chreng:
    Is that you, Rainman?

    I can only aspire to such levels of pedantry.

  • Tsaukpaetra (disco) in reply to abarker

    While I wondered about that myself, I chalked it up to a very odd case of bit-crawl. No ECC Memory for me! :crying_cat_face:

  • dkf (disco) in reply to abarker
    abarker:
    So 4 out of 5 instances of ÜberWarehouse use Ü,

    Nëêds mörë heåvÿ mætäl döts!

  • JBert (disco) in reply to Anonymous
    Anonymous:
    You probably are new here
    You're Doing It Wrong!™ It's "You must be new here"!
  • JBert (disco) in reply to abarker
    Comment held for moderation.
  • Slapout (disco) in reply to abarker

    Obviously ÛberWarehouse is a wholly owned subsidiary of ÜberWarehouse

  • abarker (disco) in reply to JBert
    Comment held for moderation.
  • aliceif (disco) in reply to abarker
    Comment held for moderation.
  • abarker (disco) in reply to aliceif
    aliceif:
    I could officialize it ...

    You'd have a fight on your hands.

  • wft (disco)

    That's why, my marketing friends, you should never underestimate researching your competition. You're going to make a laughingstock of yourselves if you brag about your software doing in 10 minutes what others do in a fraction of a second.

  • aliceif (disco) in reply to abarker

    You're a pile of clothes.

  • Luhmann (disco) in reply to abarker
    abarker:
    Apparently no one has caughtcared this yet:
    <!NotEmpty>
  • abarker (disco) in reply to aliceif
    aliceif:
    You're a pile of clothes.

    You take that back! Hats are not clothes! They are accessories!

    Also, would you really want to fight an stack of hats with infinite height? Imagine the weight when it falls on top of you.

  • abarker (disco) in reply to abarker
    abarker:
    You take that back! Hats are not clothes! They are accessories!

    Well, let me qualify that. Hats could be clothing, if worn out of necessity, such as to keep warm. But no one wears fedoras out of necessity. So clearly, in this case, we are talking about a pile of accessories.

  • mott555 (disco) in reply to abarker
    abarker:
    Also, would you really want to fight an stack of hats with infinite height? Imagine the weight when it falls on top of you.

    I'd be more worried about the effect on the Earth from having a stack of hats with infinite height (and presumably infinite mass as well) in physical contact with it. Crushed land mass causing earthquakes and volcanic eruptions, and strong orbital perturbations are likely.

  • blakeyrat (disco) in reply to abarker
    abarker:
    But no one wears fedoras out of necessity.

    Maybe they have a condition where if a woman is attracted to them, they die?

  • FragFrog (disco) in reply to RaceProUK
    RaceProUK:
    And if a random outsider can speed up your queries by a factor of 119, then, yeah, I don't want that software.
    Well, that remains somewhat to be seen. Thanks to query (result) cache, running the same report a second time can be a lot faster without any change to the underlaying database. Subsequently changing the report parameters will then once more get you a five-minute waiting time.

    I mean, I have also seen queries being sped up by a factor hundred or so just from adding the right index (or in MySQL's case, hitting it over the head with a large mallet and forcing it to use the right index manually), but in as many cases, a cache somewhere just made it look as though a query was suddenly much faster.

  • RaceProUK (disco) in reply to FragFrog

    The article stated the person added an index…

  • another_sam (disco) in reply to RaceProUK

    Post hoc ergo proptor hoc.

  • RaceProUK (disco) in reply to another_sam

    So you're saying RDBMSs don't recompile query plans when indexes change?

  • Jaime (disco) in reply to FragFrog
    FragFrog:
    Thanks to query (result) cache, running the same report a second time can be a lot faster without any change to the underlaying database.

    Careful with this. One experimental way to eliminate the cache variable is to dump the cache between runs. However, this will tell you how the query will perform in the case where the cache is totally empty, which it probably won't be in real life, so your experimental numbers won't be any more relevant. Doing the opposite - stuffing the cache by running the query before timing - has a similar problem.

    Sometimes the best answer is to see that you got some improvement in your isolated test case, then throw it into production to see how well it works with a real workload.

  • another_sam (disco) in reply to RaceProUK

    I would expect it to in most cases, but 1. RDBMSs are black magic and I don't really understand the implementation of them at a deep level, and b. if the query results or part of the query results are already cached, the recompiled query plan won't be used to fetch those cached results.

  • Jaime (disco) in reply to RaceProUK
    RaceProUK:
    So you're saying RDBMSs don't recompile query plans when indexes change?

    Kind of. Most of them mark the plan as invalid, but neither recompile them nor flush them from cache.

  • Jaime (disco) in reply to another_sam
    another_sam:
    if the query results or part of the query results are already cached, the recompiled query plan won't be used to fetch those cached results.

    It doesn't visit the cache until it is executing the plan, so any recompilation would already have happened by then.

  • flabdablet (disco) in reply to wft
    wft:
    That's why, my marketing friends, you should never underestimate researching your competition.

    And that's why, my management friends, you should never send one of your marketroids to talk to your customers' engineers.

    Of course Spencer should have known not to bring a knife to a gun fight, but that's the thing about marketroids; they're fuzzy on the difference.

  • Masaaki_Hosoi (disco) in reply to dkf
    Comment held for moderation.
  • CoyneTheDup (disco) in reply to flabdablet
    flabdablet:
    Of course Spencer should have known not to bring a knife to a gun fight, but that's the thing about marketroids; they're fuzzy on the difference.

    I think this overrates Spencer: It's more like he brought his nerf ball to the gun fight.

  • adhominis (disco) in reply to dkf
    dkf:
    Nëêds mörë heåvÿ mætäl döts!

    Except that in the case of über, the umlauts are actually the correct spelling. The word uber does not exist in German.

  • RFoxmich (disco) in reply to RaceProUK

    They're on tachyon-net.

  • RFoxmich (disco) in reply to Anonymous

    That was a far more polite reply than I would have given.

  • RFoxmich (disco) in reply to John_Imrie

    That and the fact that Spencer let him touch the guts of their product.

  • ijij (disco) in reply to abarker
    Comment held for moderation.
  • Watson (disco) in reply to mott555

    It might not even have to come into contact. Just passing by might be enough. Someone else can work out the gravitational field of an infinite rod of diameter h and the density of felt or a bit less depending on how tightly they're nested.

  • Jaime (disco) in reply to Watson

    If the stack is moving with the surface, centripetal force would be infinitely larger than gravitational force, so the behavior would entirely depend on the anchoring mechanism. I would suspect that the most likely behavior would be that the stack would fly off into space without affecting the earth.

Leave a comment on “The Upgrade”

Log In or post as a guest

Replying to comment #:

« Return to Article