I propose that the limit on aid another by specified.
Already done, under the
crafting rules. In the majority of cases, the primary crafter will have only 1 assistant.
I forsee a problem in the "fail on 1" rule. Over many days or any long project, you are eventually going to roll a 1.
It will happen. A lot. It will not be appreciated by players, regardless of how realistic it might be.
Law of averages being what it is, you're right. :/ I like the idea of having different things happen on a natural 1, though. Wolvy's idea of special stuff also fits into this pretty well. Instead of major disasters and such, I'd go with varying degrees of failure, say on a d10 -
1-2: A surge in the flows of magic occurs, and the item is drained of all magic. All gold and materials spent so far are lost, but the item can be enchanted again.
3-4: The crafter suffers a major setback. Half the gold and materials are lost, but the item's enchantment is intact and work can proceed normally. The crafter must pay half the original raw material cost again. [Basically this is the same as "fail the roll by 5 or more". I want to have a "nothing happens" result, and this is pretty much the same thing.]
5-8: The crafter suffers a minor setback. He loses 2 days of time due to problems in the item's creation. If the crafting time takes 1 week or more, he loses the entire week. [I'm going to change the original "fail by 5 or more" to "you lose 4 days out of the week" instead of the entire week.]
9-10: A surge in the flows of magic occurs, and the item's enchantment is warped slightly. The item is cursed, but the crafter does not know this. This result is otherwise treated as a minor setback.
(The above assumes that the mage has some kind of lab that cost some amount of money to build. I haven't had time to write up that part, of course.)
According to the DMG, all you need is someplace quiet, comfortable, and well-lit - anyplace that's good for preparing spells is good for making magic items, so you could do it while on the road - take a day off in the inn to work on it (which I don't agree with, personally; I think it should be done in a lab/creation room/whatever, where the crafter has access to materials and/or resources, like manuals, spellbooks, etc.).
(No, I don't have any off-the-top-of-my-head skeletons to throw up illustrating it like every one else. You want that kind of effort we have to go back to spending half a day cranking out stuff on IRC.

)
Heh.
For myself, I don't like the idea of taking 10 on a craft check and having any hope of making anything better than Average. You take 10, you're not crafting, you're cookie-cuttering. And that's what you should get. On the other hand, you take your chances and get creative, you could just produce something fantastic that you never dared hope for.
Yeah... I like that idea. Taking 10 is really for when you
know you'll make the DC, or for routine stuff that's not worth rolling for - a blacksmith churning out horseshoes, for instance. Making magic items shouldn't ever be routine (except maybe potions, and those fall under Craft (alchemy)). That gives a good justification for not letting PCs be able to take 10. As for a benefits table, just do something similar to the failure table:
1-2: The crafter makes a work of art, and somehow manages to work in a special bonus, which may or may not be known to him. The item has a minor benefit outside of its original enchantment; the DM should choose something related to the item's purpose, or related to the crafter (a cloak of resistance, for instance, made by a specialist evoker could have additional bonuses against energy; a sword made by a cleric could have an extra plus against undead; etc.).
3-4: The crafter makes a breakthrough, or discovers a faster, more efficient method of creation. He accomplishes 3 days' work for that day. If the crafting time is 1 week or more, he accomplishes two weeks' worth of work.
5-8: The flows of magic are working in the crafter's favor, and he accomplishes an extra day's worth of work. If the crafting time is 1 week or more, he gains 3 days' work.
9-10: A surge in the flows of magic occurs, and the item gains an extra minor benefit. [I can't think of anything off the top of my head right at the moment - I'm getting over a really bad headache.]
I thought of something while I was doing this, to prevent abuse/ill will from lots of really good/bad rolls, respectively: The really good/bad stuff can only be applied once. After all, once an item's cursed, it can't be cursed again, right? So only the
first natural 1 (or 20) applies as a roll on the charts; everything else is treated as "failure by 5 or more" or "success by 10 or more".