Harlequin wrote:

for example, someone wants to make a program that is a virtually impenetrable firewall. They go into a simulspace where time passes faster than normal, and spend 50 years or whatever working on it. how does that even translate in the rules? is it just a massive penalty to anyone trying to hack it? or is there not even a roll because it's so impregnable? or, for example again, if they didn't do it in a simulspace but maybe had a group of 20-30 people working on a system in real time, how would that compare and/or how long to make it that way?

If there is an efficient method to make a good product, other people are also using it. So whatever method the PCs want to use, we should assume the commercially available products are also made by that method (or something even better) with significant resources behind it. That everyday firewall around your Muse? That involved a few hundred skilled Nimbus programmers, system architects, testers and AIs working at full simspace speed for a few months. This is why I doubt anybody could make a firewall as impregnable as to prevent even a hacking roll - at most you get a heavy penalty due to the high quality of the software, but it will not be perfect. I think engineering/programming in EP is about ten times as quick (in terms of man-hours) as it is today, thanks to better technology (helpful software, downloadable libraries and modules, AI, simspace speedups, forking etc). See the discussion in this thread. This is counterbalanced by enormously more complex products (i.e. the actual engineering is likely 100 times as fast, but everything is 10 times as complex). Big software projects are big. Debian 2.2 had 55 million lines of source code and has been estimated to have required 14,005 man-years. A similarly big project in EP (like writing an entirely new AI architecture or spacecraft control software) would hence "just" require about one and a half thousand people working for a year. Single characters cannot do that. Assuming an ordinary commercial piece of software today takes about a year to develop by a staff of 100 people, in EP something of equal ability would be possible to make using just one person working for 10 years, or 10 people working for one year. Maybe a skilled person working in an area where the complexity is not overwhelming might be able to do it it themselves in just a year. See the problem? RPGs are all about heroic characters doing impressive stuff. But most impressive stuff is done by big teams, not lone inventors or engineers. If we tweak the system to allow PCs to make radical amounts of work then we should expect the NPCs to do equally radical things (even if the PCs are much better than the average person there are plenty of good NPCs around in an ordinary campaign). Taking extra time improves quality, but presumably not indefinitely. Working for 70 subjective years on a firewall means that it is likely 70 years out of date - those hackers have also been working, and the perfected firewall will likely be good merely because it is so unusual and incompatible, not because it is clever against the threats that exist. Even maintaining the will to run an endless project is tough - companies sustain it by paying employees, open source by social rewards. But a PC better manage his willpower rolls. However, in order to get a game system anyway, what about this: every doubling of time gives you another +10 on the quality?

I am just going to sneak in here and point people towards the Elite Exploits sidebar on page 246, which gives a guideline for improving software using Programming. Quite clear how it works. Can be done on Firewalls, Exploit programs, maybe even Tac-Nets.

The Eclipse Phase model seems to be "spend X resources for +Y to a given roll", and the "Elite Exploits" sidebar follows this model. Programmer time is traded for equipment bonuses to software; exactly how that works isn't necessarily defined, and if my players wanted more detail, I'd encourage them to supply some - and give them a narrative bonus if it sounded good. So the things you can do with Programming boil down to a few categories. You can make a program that gives a skill bonus to someone, or penalizes someone else trying to do bad things to you (example: "classify whoever I'm looking at using the Facial Action Coding System and run this expert system on that data, giving me +10 to Kinesics when face-to-face with someone"). You can make a program that does something a rather stupid transhuman could do, only repetitively and quickly (example: "watch these public spimes for my face, then send the email I wrote earlier to this address when you see me"). You can program something that takes input and does something with it, or does something and sends output about it (example: "gather data on official habitat population by mesh use, then gather information on oxygen usage and tabulate discrepancies").