Leaving Oracle | Hacker News Reader

Leaving Oracle

dtrace.org

64 pointskmavm23 comments

I seriously wonder what the Oracle upper management thinks about the recent exodus of their (actually, Sun's) top talent. Do they even notice it? Do they consider it as a problem, or just "getting rid of them ballerinas"? Or maybe they think that they are easily replaceable?

If Oracle is seriously planning to "spend more on Solaris than Sun did", they have a lot of hiring to do... just to get back up to parity.

At least in Austin it's rumored that Oracle has a net influx of chip designers.

Sun had a ton of chip designers, I'm skeptical quantity was the problem with respect to their SPARC development.

Someone reliable said that one of the things happening is that a lot of Sun software engineers had their salaries significantly reduced in the transition. This would suggest that either Oracle plans on having different people work on Solaris (and as you note many of them) or they were lying through their teeth.

Well see, you can spend money on engineers to improve the product, or you can spend money on strippers and steaks to sell the product.

Technically, both are "money spent on the product" in some way, shape or form. Oracle are just more likely to throw money at the latter class rather than the former class.

Why does Oracle want to be in the chip design business?

Because they want to sell complete systems that provide more power than the x86 ecosystem can?

This has been said before but it's worth repeating, Oracle is not an engineering company. They are giant sales organization with an engineering department tacked on.

They probably considered Sun's workforce before the acquisition as a liability instead of as an asset. The assets were in Sun's products, patents and customers. In that calculus, the people who quit willingly just represent severance checks Oracle doesn't have to write.

But won't the products suffer severely once the brains behind them leave? No organization as old and large as Oracle can discount that.

What the grandparent post said is correct -- Oracle is a sales company. They sell a seemingly engineering product -- a database. But that doesn't matter. Most organization who buy that database could probably manage with Postgres if they had to (not saying "all", just "most").

So why do they buy it? Because Oracle salesmen play golf well and know how to lose to the right CEO/CIO/CTO person. Yes, Oracle DB is "golfware" not "software".

There is a whole ecosystem of training and support that they have built. Keeping that up is worth a lot more than rolling out better software, because that is where the money is coming from.

So when the best Sun engineers leave, Oracle managers are probably thinking "good riddance, we could replace them with 10 engineers in India for the same price!" and the sad thing is, -- they are probably correct for the market they are in.

upvoted for "golfware"

Thing is, there was a long period in the '90s when Oracle was the best solution, period. Non-mainframe DB2 was by the mid-90s (and perhaps earlier) nice, cheaper and simpler to administer but used conventional locking instead of MVCC (works for a small group of DB engineers but doesn't scale as well with lots of developers), MySQL was a non-ACID nightmare (heck, unless my research was really off in mid-1997 Linux didn't truly impliment fsync back then), PostgresSQL wasn't ready for prime time and the less said about SQLServer the better (it's improved since them or so I'm sold). And small fry like Informix weren't particularly worth dealing with, Ingres was dying, etc. etc.

You can see how the above golfware paradigm could easily sink in when you really did have arguably the best solution, even if it cost twice as much as the next best (DB2, who's ecosystem (e.g. books) was and I'm sure still is much less rich).

However as you and many others have noted the world is very different now.

I remember the same situation around those days: everyone wanted to get a database behind their site, but MySQL felt like you were running with scissors. It was easy to sell Oracle to clients then despite the $5k+ cost. Oracle's sales cycle may been protecting them until now, but they must realize what will happen when they abandon their strong technical underpinnings.

    They are giant sales organization with an engineering
    department tacked on.

Years ago we had oracle sales present to our office. During questions I asked what was being done to make sqlplus feature competitive against the free software consoles like psql, mysql on basic features (e.g. browsable command-line history). The presenter was surprised I had complaints about it. He fired it up on his laptop on the projector. It crashed during startup. Laughs all-round and beyond that the Oracle reps didn't even pretend to be interested. I'm not sure who they thought they were selling to.

I came to the same conclusion as you.

In the early 00s I was at an Oracle sales conference. Now I was never in Sales myself mind, I was a consultant (and actually employed by an Oracle partner) but the sales types always had the best canapes and cocktails, so we blagged our way in.

So, it's time for the main event, in the auditorium. They got smoke machines. They got lasers. They got Vanessa Mae (or a clone thereof) wearing next to nothing and playing up a storm on the electric violin. They got cheerleaders and they got MCs on stage rappin' about bringing Oracle's rivals TO THEIR KNEES. And the crowd (99% Oracle salesmen) went wild. I'm at the back wth my little clique of weapons-grade cynics and we're like, WTF have we gotten ourselves into.

Oracle ain't even a sales organization, they're a full on religious revival.

Adam is a seriously talented developer. I had the privilege of meeting him, as I'm sure many have over the last several years, at a Linux conference in Los Angeles a few years ago. His dtrace presentation still sticks out in my mind as a superior demonstration of superior software. Whomever ends up with him on their staff will be better off for it.

It looks like the Valley is about to get a serious influx of unburdened creativity and talent able to get working on the next big thing. Oracle's loss may well be lots of other companies' (and FOSS in general I hope)'s gain.

I would imagine (or at least I hope) that Oracle was fully aware when they bought Sun that their company philosophy was going to clash with Sun's and they would have lost a lot of bright people. It's a pity, but I don't think it came as a surprise to them.

I think they were well aware of this. The goal of the management is to remove the sun culture and replace it with Oracle culture as soon as possible.

The hardcore believers who are leaving simply helps Oracle.

Speaking of Oracle, that easily leads me to the discussion of OpenSolaris (RIP). Check out www.illumos.org

It's is starting to get quite interesting, also check out #illumos on freenode.

I wonder how many patents Oracle has on this one.

The curious part of me was hoping for a reason.

I can't imagine what reason there could possibly be.

Reply on news.ycombinator.com