HP goes bundle mad with quicker HP-UX

i mean, really, does anybody seriously use HP-UX? From the ugly, V7 "make" to their odd C compilers, to the poorly-documented "segment register", HP-UX is a throwback to the Bad Old Days of circa 1990, with BSD vs SysV conflicts, and the "unix wars" that were probably provoked by MSFT/Wagg-Ed agents provocateur.

HP-UX: just say "no".

My former company uses HP-UX and AIX machines. They're all old though. You know the kind: The really massive mainframes. But they still work and the applications on top of the do too. I strongly doubt that they would buy new HP-UX or AIX machines though...

I work with AIX quite a bit. Recent (5.x) versions are a lot better than the older versions, but there are still things about it that I don't like. Many of the out-of-the-box deficiencies can be corrected if you download, compile and install a bunch more open source stuff to make it even more Linux-like. Getting generic open source to build on AIX is much easier than it used to be. The only downside to that is that it complicates AIX upgrades as you have to go through a lot more work to update everything else that doesn't come in the box. Also you tend to build dependence on those extra libraries and utilities into code that you develop, which adds a lot of pre-requisites when you need to install on another AIX box.

As for HP/UX... I've not worked with really recent versions, but it used to be comparable to or worse than the older versions of AIX in terms of wonkiness and portability issues. I don't know if HP has been as diligent as IBM in correcting that or not.

FWIW, if it were up to me, and if some of the applications I work on didn't have some weird dependencies on a few odd proprietary bits of AIX that would be inconvenient to work around, I'd gladly replace the AIX boxes around with Linux...

I don have fond memories of the old HP compilers, though.

Sometimes it seemed like they would let anything through unless you added something along the lines of "THIS IS VERY, VERY WRONG: NO C CODE LIKE THIS SHOULD EVER EVER COMPILE SO PLEASE, RAISE AN ERROR SO THAT I DON'T GO NUTS FINDING OUT THAT I MADE A DUMBO MISTAKE THAT YOU SHOULD HAVE CAUGHT!!!!!"

Wow, a useful error message. Beats the heck of out the typical "the process exited with return code square root of pi. What does it mean? That's for us to know and you to find out."

Yes, the HP-UX printf() is known for overwriting a format string with output.

My previous employer (2 years ago) developed for HP-UX, on both PA-RISC and Itanium. Several potential customers had come forward, asking for our data security product on HP-UX. I was directly involved in getting our product certified by H-P to run in a clustered environment.

My employer decided to standardize on hpux back in 1997, though I had recommended sun. hpux seems to be favored by stodgy, risk-averse, change-averse organizations.

Fast forward to today: we're running terabytes of oracle databases on hp-ux, and... preparing to migrate it all to linux. heh, I'd been telling them for years, that if oracle on linux works so well for amazon.com, it will work for us... finally, the penny drops!

Doesn't Oracle develop Oracle on Linux, then port it to other architectures?

http://www.builderau.com.au/architect/database/soa/Oracle-to...

> Doesn't Oracle develop Oracle on Linux, then port it to other architectures?

Yes, according to the oracle people who came to talk to us a few weeks ago.