5 ms·

The same story told again -- it's aim is to generate this magical atmosphere around a fast db engine and language that's deliberately obfuscated to make people

The same story told again -- it's aim is to generate this magical atmosphere around a fast db engine and language that's deliberately obfuscated to make people who work in it feel smart, so they try to spread that it's the best. KDB/Q is a nice tool, not the holy grail how they put it. And it's not fast because they know something better - it's because it lacks almost any safety measure.

sohagan857 12y ago
Ridiculous comment. Have you at least tried q? K? Kdb+?
- avmich 12y ago
  Yes, interesting. APL was meant to be used for teaching programming - with it's exceptional clarity (sic). Java took the place... J engine is claimed to not have memory leaks - that's from very early versions and being written in C. It's true that it could be hard to think about the problems "vectorally", but still - I agree, the comment sounds ridiculous.
- sz4kerto 12y ago
  I've worked with and later run the kdb/q stack for a couple of years at a financial services firm. I certainly enjoyed it but I've also seen the downsides of it, and according to my experience it does not worth the investment, and the biggest reason behind its adoption is A.W.'s relations with the financial industry. By the way, I really don't want to go into anything personal but your comments represent the attitude around kdb well.
  - qfan 12y ago
    It is a fun environment and the community is very friendly and smart too. Its widespread adoption in finance is because it is a proven tool for processing the huge datasets that other systems cannot. If you have a business case for that, clearly it can be a worthwhile investment. Other industries are beginning to see value in leveraging their data sets and that's probably why after being in the business for 20+ years kx continues to expand their customer base; note they now have pharma and energy customers. Yes, there are difficulties involved in managing huge data sets, but IMO no other tools does it better than kdb+.
    - kodisha 12y ago
      can someone please give example on how the pharma industry is using this?
      - sohagan857 12y ago
        This might be what you're looking for..http://kx.com/_papers/Kx_Pharma-Vertical_Markets-140502.pdf http://kx.com/_papers/Kx_Pharma-Vertical_Markets-140502.pdf
kiyoto 12y ago
I totally agree. That's why I wrote a parser for the language to help me and my coworkers debug =) https://github.com/kiyoto/ungod https://github.com/kiyoto/ungod As I commented above, the syntax is horrible. But I still think it is an effective tool for certain problems.
MichaelGG 12y ago
I'm very interested in hearing more. What do you mean by lacking safety? As in the DB invariants aren't? Or lack of durability? Or as in, easy-to-misuse query language? And to help calibrate, what are your preferred languages and styles?

sohagan857 12y ago

Ridiculous comment. Have you at least tried q? K? Kdb+?

avmich 12y ago
Yes, interesting. APL was meant to be used for teaching programming - with it's exceptional clarity (sic). Java took the place... J engine is claimed to not have memory leaks - that's from very early versions and being written in C. It's true that it could be hard to think about the problems "vectorally", but still - I agree, the comment sounds ridiculous.
sz4kerto 12y ago
I've worked with and later run the kdb/q stack for a couple of years at a financial services firm. I certainly enjoyed it but I've also seen the downsides of it, and according to my experience it does not worth the investment, and the biggest reason behind its adoption is A.W.'s relations with the financial industry. By the way, I really don't want to go into anything personal but your comments represent the attitude around kdb well.
- qfan 12y ago
  It is a fun environment and the community is very friendly and smart too. Its widespread adoption in finance is because it is a proven tool for processing the huge datasets that other systems cannot. If you have a business case for that, clearly it can be a worthwhile investment. Other industries are beginning to see value in leveraging their data sets and that's probably why after being in the business for 20+ years kx continues to expand their customer base; note they now have pharma and energy customers. Yes, there are difficulties involved in managing huge data sets, but IMO no other tools does it better than kdb+.
  - kodisha 12y ago
    can someone please give example on how the pharma industry is using this?
    - sohagan857 12y ago
      This might be what you're looking for..http://kx.com/_papers/Kx_Pharma-Vertical_Markets-140502.pdf http://kx.com/_papers/Kx_Pharma-Vertical_Markets-140502.pdf

kiyoto 12y ago

I totally agree. That's why I wrote a parser for the language to help me and my coworkers debug =) https://github.com/kiyoto/ungod https://github.com/kiyoto/ungod As I commented above, the syntax is horrible. But I still think it is an effective tool for certain problems.

MichaelGG 12y ago

I'm very interested in hearing more. What do you mean by lacking safety? As in the DB invariants aren't? Or lack of durability? Or as in, easy-to-misuse query language? And to help calibrate, what are your preferred languages and styles?