QUOTE (Prime Mover @ Oct 22 2008, 10:05 AM)

Now onto "Why can't I make/program it, I have a 7 skill, run a cracker gang and am worth several hundred K?" Well other then interval time constraints whats a GM to do?
When I read this in the core book:
QUOTE
MATRIX ATTRIBUTES
Matrix attributes generally range in scale from 1 to 6, with the
lower ratings indicating cheap, outdated, or salvaged components/
soft ware and higher ratings refl ecting well-made parts/code. Some
cutting-edge and prototype models may exceed rating 6 attributes,
but these are exceptionally rare and hard to come by.
I see little problem in having it exist, and no problem adding some house rules adding requirements for it's development or aquisition.
Looking through Unwired, the only two places I remember having hardware higher than 6 were an Azzie facility and a Zurich Orbital ground station, both of which is going to have multi-million, if not billion nuyen, IT and development budgets.
A fully decked out commlink at 6, before programs, costs a little under 17000, with all rating 6 programs (not counting the new ones in Unwired) it jumps to just under 100k

. That's at mass market pricing (or as close to mass market as high end and hacking software gets), taking full production runs for hardware, volume sales, etc. for reducing the end user cost. That isn't going to happen when it's prototypes or custom-built hardware.
I see it something like this, a console game these days can run more than 2 million dollars to make, so at $60, even if it was pure profit, would need to sell over 30,000 units to gross 2 million, and the game probably wouldn't be considered a rating 7+ program. If a hacker has a million nuyen to burn, I could see a possibility of being able to buy rating 7 hardware and some programs to make full use of it.
Since 1-6 covers what is fairly available on the market (legit or pirated) for hardware and software, I'd say if you want something at 7+, I see 3 straightforward ways you could go about getting it: (to reduce the size of the post I'm putting my thoughts on each in spoiler tags)
1. Raid on a Corp development/research lab.
[ Spoiler ]
I would say it has to be a B rating for a corp that specialized in what you are looking to get, A rated or better if they aren't. They are likely to come after you since they've (likely) spent a couple million on development/aquisition of it to begin with. A corprate strike team or a couple of shadowrunning groups coming after you could make for a decent couple sessions as well.
2. Have a high rated contact in the right corp/shadow group (at least 5 connection rating) and you pay black market prices for it (the guy's most likely risking his life to get it to you).
[ Spoiler ]
I'd throw fencing modifiers in and a multiplier of x20 for rating 7, x50 for rating 8, and x100 for rating 9 (deltaware is considered device rating 6 and has a multiplier of x10).
Say a commlink at all rating 7, if available for mass market would cost 200k (matching rough increase in SR4, response doubles, signal x3.5 previous rating, program multiplier x2.5 (500 becomes 1250 per rating point for 7-9))
With +20% for Distribution Channels Monopolized and the x20 mutiplier, that's 4.8 million

. It's not going to be cheap. Having the contact report it stolen (-20%) to knock it back 4 million. If you want to be nice, forgo the increase but still have it reported stolen for a price of 3.2 million

. Your paying a premium price since, outside corp labs, there may only be a few hundred at that level in existance.
3. Find a way to write it yourself.
[ Spoiler ]
To even consider writing anything at rating 7, you'd have to have Software at 7, and to get to 8 or 9, you have to have a Software specialization in that specific program (instead of the type like Offensive, Special, Defensive etc.). Doubling the normal interval (1 month to 2 for most programs) should cover the added complexity of having to utilize new machine instructions, additional pipelines, etc. I'd also reduce the effective rating of any Programming Suites used by 2 to reflect that they were not made for programming on the protoype harware. If you're developing on hardware that can't fully test it, add a 2 die penalty (i.e. Rating 7 program without Optimization written on a System 5 commlink would have a -4 penalty). Then you have to consider that the SOTA continues to move forward while you develop yours, so the Threshold could continue to go up if you can't develop it quickly.
For a quick example of what I would consider:
Hacker with Logic 6, Software 7(Decrypt +2) writing Decrypt 9 on a System 6, to be used on a prototype, Extended Test (Software+Logic, 18, 2 Months):
Logic 6
Software +9
Programming Suite (4) +2
Prog. Environ. Access Interval/2 (I don't like dividing the interval, instead I would give 2 tests per interval)
Inadaquate Tools -2 (prototype hardware rarely has fully developed software kits or compilers)
Superior Environment +1
Underpowered System -6
That give 10 dice to get 18 hits, assuming no glitches, for an average development time of 6-9 months (12-18 without the Programming Evinronment). Another option would be to forgo the interval increase but require the Programming Environment Acess and remove the interval reducing effect. This assumes the Threshold never changes. If you wanted to get nasty, up it every roll by 1 or 2 (more if there is a glitch). It would almost guarantee you have to have a software team to get it done.
Upkeep/Patching would only need 2 hits, but at a base 2 week (or 1 week) interval, so it would be very difficult to keep up more than 1 program and still be available for runs.
Under the current rules, it would be (Soft+Logic, 18, 1/2 Month) and 19 dice to do it under the above example (without house rule modifications), giving an average development time of 1-2 months, which just seems to short to me. Add in the scaling Threshold and its 2-4 months, and add the underpowered system house rule on top of that, it bumps it back out to around 7-10 months; which is still decidedly quick for 1 person to code the equivalent of Photoshop, Word, and movie editing software all in one, even for someone with genius level intelligence.
Lastly, don't forget that new items/programs are always being developed, so that SOTA response chip you put in today could be the norm 6 months later. I'd degrade the rating on Response by 1 every 6 or 12 months of gametime until it's at 6 to reflect new hardware released.
On a slightly related note, as a house rule, I would apply a rating to Lifestyles (1 for Street, 6 for Luxury), average (rounding to nearest) which ones you were at after a year in game, and if you have any legal programs or commlink ratings above the average, you have to pay to keep them there.
[ Spoiler ]
This would be to cover hardware and software upgrades not covered by your lifestyle. For example, a hacker has a Street lifestyle for a year, but has 5 rating 6 hacking linked to a fake license (to avoid having to maintain them himself) and a Rating 6 commlink. To keep them from degrading 1 point each for the next year, he's got to pay the upgrade cost for the next engine, new response chip, etc. since his lifestyle only covers keeping them at Rating 1. Paying for 1 rating point for each program (from 5 to 6) and a new processor (Response 6), it adds a once a year upkeep of 16000

(8k for Response 6, 5k for hacking programs, 500 for System, 500 for Firewall). This would be the same if the hacker had a Middle lifestyle as well, but at a Luxury level, it would be included in the cost of living. I think something like this can balance the ammo costs of a sammie, or binding material costs of a mage. I also assume the cost of living covers getting a "clean" commlink regularly attached to whatever identity is attached to the lifestyle, and a mage or sammie doesn't have to bother speccing a commlink if they don't want too. Instead, they just have a commlink at whatever their current lifestyle rating is, loaded with all common use programs at that rating (so a middle lifestyle gets you a 4/4/4/4 commlink with rating 4 programs).
Using something like this would give you a way to calculate the cost for keeping a rating 7 system and programs at 7, once you developed a cost structure.
As a quick touch on the Missions discussion, to probe the node you make a Exploit+Hacking(16, 1 hour) extended test (assuming your in VR) just to get basic access (the threshold is 22 if you want admin rights). If you average 4 hits, that's 4-7 hours to break into the node, as apposed to on the fly at (10, 1 IP) or (16, 1 IP) for admin (about 2 combat turns). Depending on how much time the rest of the mission has already taken, you may not have time to probe the target, so you risk the system's Analyze+Firewall (Stealth, 1 IP) test to detect your hack (which would likely start a trace if you triggered an alert). Given the nature of the module, limited play time, and with the given information, I don't see why this wouldn't be a matrix accessible node for the Denver Ares corp office (I think that's were that mission takes place). Heck, given the description given in previous posts, I wouldn't even have IC or spiders reveal themselves until a trace was completed, if then; just backhack the commlink to provide a location and guide KE in for the capture.