So what exactly can I find with data search and how much do I need to know to start?
If I have a picture and voice recording of my johnson, can I discover his real identity or do I have to actually hack into his corp's system to get the info?
If I have a die pool of 12 (not impossible for a minmaxed character.) and therefore on an extended test can roll 144 dice, thats going to come out to 36 bought success or 46 successes on average. Does that mean i pretty much automatically know anything that I think of to ask?

It's like googling: if the information is publicly available online, you can find it...it's just a matter of time, and sometimes skill. By the same token, the Data Search skill shows how quickly you can find the data you need in a given database or mainframe.

Let's say, just for shits and giggles, you are a hacker looking to find out Mr. Johnson's identity. You have a Data Search skill of 3 (Speciality: Corporate) and a Rating 3 Browse Program, so you're rolling 8 dice. The GM rules that the difficulty is hard - it could be much more difficult, but Mr. Johnson is new in the game and put his picture up on a Matrix dating site. You're searching the entire Matrix, so the interval is 1 minute.

If you roll two hits per attempt, you meet the threshold after 4 tries - in 4 minutes of dedicated searching, you can find Mr. Johnson's dating profile. The character can go on with further searches from there.

Now, it may seem that data searches are guaranteed for success, but this isn't the case. The GM can set the threshold, or add dice pool modifiers, as they wish. The number of rolls you can make on the Extended Test may be limited by your dice pool, or by how much time you have available. Certain information - like what color panties the president is wearing today - may simply not be available, though a data search might reveal what host you need to hack to find out.

Of course, a min-maxed character will be able to find out almost anything publicly available, but that's the benefit of min-maxing.

I did this exact test with my Technomancer, we had a Mr. J that didn't pay us so we were going to extract our nuyen. A simple Matrix-wide data search turned up nothing. One of the other characters figured out through a contact what area of Seattle Mr. J was staying in. Then I started hacking local nodes that had extensive video surveillance of the area I was looking in. Several nodes later I found Mr. J going into his hotel.

So Data Search is more useful if you know where to look. That "full Matrix search" does not enter any SANs.

While I feel a GM shouldn't do this often, it would be possible for the GM to determine a hard limit on the info you can get, if anything at all, from a matrix search.

I'd say it'd become harder if the target you're searching is either a hacker, or regularly employs hackers to spoof his own trail/online info. I'd also say it's possible that your target of the datasearch could take countermeasures to that in the form of agents/whatever that are online checking to see who's checking up on him. So low level datasearches might not get the whammy, but if his agents/security hackers/whatever notice someone is spending a concentrated effort trying to hunt down flagged information, that may have consequences just as you were trying to hack a node.

'Data Searching' seems to be a more common level skill as opposed to hackermindset, as such I'd say its a little more overt than a hacker trying to covertly pick up information. It'd help a hacker know where to start, but for their own security, they'd be better off switching to hacker-mode as soon as they could to minimize their own trail.

Also, consider that a lot of information on the matrix might be missleading, outdated or simply be wrong (by design or otherwise ).

Just like today .

Yeah, I was just reading a story about that the otherday, MSNBC I think, one of the reporters did an ID check on himself, and found out apparently he's a child molester. Well, not him, but a guy with his same name. But the service still sorta implied it was him. Scary stuff.

so this came up in the context of the last game I ran. A player pointed out that irl, a data search was run on general wesly clark, got all his credit info, his cell phone number, the lst 100 calls on his cell phone etc, dreckcetera.

In RL the internet is just burgeoning and politicians are writing incriminating emails that get them is trouble later because they don't realize that every email is recorded a few times by various servers.
In Shadowrun, at least for game balance purposes, IMO it should be assumed that every person who could consider themselves any type of target is Matrix-saavy or a least hired a Matrix-saavy person to make sure that any assiociated SINs, DNA markers, Fingerprints, personal vehicle commcodes or any other sensitive information is safe from Hackers.

Guys... if your average hacker can Data Search to figure out everything he wants to know about Mr. J, then Mr. J can hire your average hacker to go find out what the Matrix knows about him and have it erased or modified.

I run Data Search in a very flexible manner... you don't just find out "more and more" data with better search results, you start getting more accurate. I might decide that up to 3 successes on a test tells you some stuff about the target, but 4+ indicate that some of the things you may have found with less are actually incorrect or misinformation.

just picked up "On the Run" and I think you need to up those numbers, 16 successes is about the equivelant with a leg work/knowledge skill of 4 successes. This is generally the highest level of successes available and anything beyond that is flair.
Also parrt of my point is that with 25 bought successes which is what your decent hacker (skill group 4 browse program 6) is gonna have, he's going to make knowledge skills and most legwork useless.

*shrug* Sure, as long as that data's in publicly accessible databases or has been on the news. I just threw out those numbers as an example... it's up to the GM to set them wherever he wants.