I don't, and I'd said as much in a previous post... and I agree that there's an ego boost aspect. However, I don't think these rules really accomplish this. As I pointed out, there are core spells that pretty much render mundane crafting pointless.
If you purchase a Masterwork Longsword, you spend 315 gp. If you buy a standard longsword for 15 gp, the Masterwork Conversion spell costs 300 gp. If you Craft a masterwork longsword, you invest in materials costing 105 gp. If you fail a roll by 4 or less, you make no progress. If you fail by 5 or more, you need to invest another 1/6 of the item value.
So, a sword. Spend 5 gp, Take 10 for a 20 and finish it in about 3 days. Now we need to Masterwork it - that will take another 7.5 weeks of Taking Ten. What if you only have, say, +5 (you're first level with a +1 INT bonus and one rank in a class skill, working with standard tools)? You need to roll a 15 to succeed, and an 11 to avoid more cash in, so half your rolls will cost money - you're not skilled enough to make this process effective. So what? You have limited natural aptitude and limited training, so maybe you're not a master craftsman yet.
Once you can do this reliably (same INT bonus, 3 more ranks and masterwork tools, say), you can reliably craft masterwork items - it will take longer, but save 2/3 of the gold.
Also, that sword you made... unless you're using rules for your weapons to scale with you, you'll be hanging it on the wall never to be used again (or more likely, sold) because you have a better one. (And yes, I know not everyone's game is like this... but the core rules are, and that's what this thread was initially about.)
Based on the RAW, I can craft a Sword at 1st level, Craft it up to Masterwork (as a separate element) once I can reliably make that check, and then enchant it (or take Master Craftsman as a feat if I want to Craft magical items without being a spellcaster). So I can enhance that specific sword as I go along.
I don't think that Pathfinder's redesign of magical crafting for 3.x is much of an improvement. In 3.5, the crafter had to spend XP, which pretty much meant that they never crafted, or if they did, only certain select items for themselves.
So what result do we want? A system that no one uses is useless. The RAW, which you are appropriately using as a baseline, assumes a thriving market in magical items, as you are pretty much assumed to be able to buy whatever you want to enhance your abilities, not work with what you find. Crafting cannot, then, be all that difficult or risky, or that marketplace would not be there.
The problem with that is that the spellcraft roll is, for the most part, trivial... and in cases where there's a significant chance of failure, most players won't take the chance because a) most players are risk averse when it comes to gold or XP, and b) the failure (barring a GM really investing something into it) isn't interesting.
And this is different from other PC abilities how, exactly? They aren't going to craft items they don't have the skill to craft. Even failing 2 rolls by 5 still leaves them ahead on a cash basis, and a failure slows, but does not end, progress. Handle Animal can fail, but it doesn't kill the animal or make it untrainable - it just means it will take longer to train it.
Perhaps the better answer is that we should have a standard "item crafting" mechanism that provides, say, a 50% chance of success, a 25% chance of making no progress and a 25% chance of a very poor result (say, loss of investment or a cursed item, maybe the curse only has a 5% chance) for a level-appropriate item. This system would be automatically available to all spellcasters with no investment of character resources. But if you want to be skilled at crafting magic items, you take the appropriate Craft feat. This allows you to Take 10 on the roll, and perhaps bring the time down to the present level from a higher time investment under the "no feat" option.
Now crafting magic items would be more like crafting mundane items - pretty much anyone can try it, but you need to invest some character resources to be decent at it. Seems a lot like combat maneuvers - anyone can try them, but you won't likely be very happy with the results unless you take a Feat or two in order to be really good at any one maneuver.
If the ability to craft magical items is of limited utility, with a significant chance of failure, so there's a low incentive to even try, this implies two things. First, PC's are asked to spend character resources to move a task from "you can't do this" to "you suck at this" if we still require they take feats to be permitted access to this high risk option. Second, we need to rethink the ready availability of magic as a commodity - it's clearly not a safe or profitable endeavour, so there's not likely to be a thriving magic item manufacturing economy.
Remember, this is a world where Detect Magic is a level 0 at will, and you only need to make a spellcraft check to identify something. The crafter will immediately know it's a cursed item, and that's the end of that.
Why should there be a lot of cursed items? What are the chances of running across a mundane item which is so shoddily constructed that you take a penalty when using it? I'd rather see "cursed" items that have both benefits and drawbacks, so the players have to consider whether their use is actually worthwhile, perhaps the occasional specially crafted trap (I recall an old module with a false treasure room chock full of cursed items) and maybe the occasional uniquely cursed item.
