Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror

Comment Re:Whoops (Score -1) 510

There are some benefits to renting even if more expensive long run:
1. It can be easier to fully deduct the rental as an ongoing operation expense.  A computer purchase might have to be treated as a capital expense and probably depreciated over serveral years.
2. Depending on the terms of your contract, you might not be fully liable if a disaster (hurricane, tornado, flood) destroys your rented computers.
3. Again, depending on the terms of your contract, if the BSA or sheriff's office raids you and takes the rented computers, it might not be your problem, you rent another batch and keep on working.

Comment Re:Ha! (Score -1) 117

In my previous job I had to support customers with multi-petabyte databases on SQL Server 2008 and Oracle 10R2/11R2.  SQL Server is significantly easier to install and manage.  The tools to perform administrative operations such as creating new databases, doing database backups, profiling database server activity, measuring performance are much easier to work with on SQL Server.

SQL Server pricing also works out to less than half what Oracle charges per socket.  Last time I priced it out, Oracle came to around $20,000 per socket and SQL Server was under $10,000 per socket.

IMO the only advantage that Oracle has is better high end scalability on very large database servers: multi-petabyte, > 64 core and > 128 spindles/flash drives.

Supposedly SQL Server 2010 has significantly narrowed that scalability gap, but I left the company before they switched.

What I recommend to my customers is:
Databases with small transaction volumes and few concurrent connections: MySQL, PostgreSQL, or SQL Server free edition
Medium databases with moderate transaction volumes (10GB to 1TB): PostgreSQL or SQL Server
Large database with high transaction volumes (1TB to 100TB): PostgreSQL (if you have inhouse DBA with optimization experience) or SQL Server
Very large database (100TB to 10PT): SQL Server if performance requirements can be met with mid-sized hardware (up to 8 sockets, 64 cores), otherwise Oracle
Girnourmous databases (>10PT): Oracle if performance requirements can be met with large-sized hardware (8 to 32 sockets) or using the extremely costly Oracle Exadata custom database hardware stack, otherwise one of the specialty databases such as Teradata

Comment Re:Ormandy did excercise responsible disclosure (Score -1) 497

Your systems would be statistically safer.  The probability of any given XP system being attacked by this exploit after the disclosure is much higher than the probability before the disclosure.

Yes, this exploit was probably being used by a limited number of people before hand, but now it is being used by 10x as many script kiddies.  Overall, systems running XP are much worse off because of what Ormandy did.

Comment Re:Mistake my ass. (Score 5, Informative) 479

I worked for a while for a company that makes software for modern slot machines.  Each state and indian reservation has different rules, but in the one we wrote software for it works almost like a roll of scratch-off lottery tickets.

When developing a new game, the company decides on the payout, for example, 95%, which means that on average, out of every $1 played, the company pays out 95 cents back to the players.  The company then decides on the prize distribution, for example (not a real game distribution, just an example):
Count   Prize   Payout Amount
387,251 0       0
10,000  1       10,000
2,000   5       10,000
500     25      12,500
200     200     40,000
30      1,000   30,000
15      5,000   75,000
3       30,000  90,000
1       100,000 100,000

400,000 136,231 367,500 Total

So out of 400,000 games played of $1 each, the casino is paying out $367,500 and making $32,500 profit.  The prizes are randomized and the resulting distribution inspected to make it is distributed appropriately.

The prize distribution is saved in a central casino database.  Every time a play happens, while the graphics or reels are moving, the machine talks to the central server over a secure network and requests the next available prize.  The server finds the next prize in the the list, marks it as played, and sends it to the machine.  If it is a win, lights flash, bells ring, etc.

Casinos in general want big jackpots, as loud and as attention getting as possible, since it gets more players to play longer.  They have no interest in cheating you out out of big prize, since they are making money on average every time you play.  Their interest is to keep you putting in money into the machine as long as possible, and they do that by having jackpots as often as they have calculated they should do it.

Slashdot Top Deals

Byte your tongue.

Working...