spongefrob

Hosting and rackspace matter a WHOLE lot if computers are central to your business. Would a bank get started without a vault? Can you build a house without a foundation?

The point is as cheap as possible... not as cheap as cheap...

If you're going to end up with a lot of servers: get into rackspace NOW and develop a relationship with the datacenter. If computers are at the center of your business then you owe it to your business to house them with the same care you would your employees.

Or, you could grow it ad-hoc and deal with serious downtime and expansion issues later and attendant costs. Choice is yours. I would consider how well you like stress before deciding.

Convincing people to use their own computers is epic fail if you have intellectual property issues (or think you might in the future.) If your employee is going to sign an NDA under no circumstances should you allow them to use their own computers.

Also, varying degrees of attention are paid to things like vulnerabilities and stability as well as standards. If your relying upon yourself to do IT, then you're automatically setting yourself up for connecting ten different computers in 15 different ways. This does not scale.

Then of course, you'll have the hotshot programmer who'll set up his own sourcecode repository and you'll have no end of headaches getting the two repositories to agree. If you're going to rely on computers with minimal IT staff then you need to go cookie-cutter approach in the beginning and set up definitive rules like "a server is a server" and "a desktop is not a server".

Quite. As I said, Apple, in terms of quality, is in the top three. I would rate Samsung in the top one...

Also, if you're not permanently ensconced in your present office space, having space in a data center means a move without moving your servers (except, of course, for the network portion) It is as close to painless move as you can get...

I've been at jobs with small monitors and crappy chairs and I've been at jobs with small monitors and good chairs. I've also been at jobs with good large monitors and crappy chairs. If you offered me the choice of an Aeron Chair and a 30" cinema display, I'll take the chair. For software and network engineering related tasks, the single biggest enhancement of productivity is a good comfortable chair. The Aeron chair fits this bill.

But that doesn't, per se, mean that a good solid monitor isn't also a help. After 20 years in IT, I'd rate Apple in the top three for monitor quality.

If you are going to have servers, backups and other items that are both mission critical and environmentally sensitive then you need space that is air conditioned, expandable, properly sized for power and lockable. Unless you are both a licensed electrician and a trained HVAC engineer I guarantee you that whatever assumptions you have about servers and their power loads, heat signatures and space requirements are wrong.

Unless you are planning on hiring full time IT staff, you probably don't want to build your own datacenter. If you think you are going to get much much bigger (server wise) then you might want to look into buying a small (but expandable) amount of rackspace in a datacenter. This will place all server infrastructure issues at somebody elses feet.

If you think that you'll buy three or four servers and not expand beyond that for the life of the business, it's probably a good idea to simply modify a large closet (> 150 sf) for AC, UPS, power, etc... But if you think you are ever going to have more servers then that you need to think about expansion now.

If the servers are critical to your business then you'll thank me later...

<blockquote>If we say we're looking for x, we'll get applications proposing x, certainly. But then it actually becomes harder to judge them: is this group proposing x because they were already thinking about it, or because they know that's what we want to hear?</blockquote>

Why is this important? I don't recall reading anywhere that Y Combinator takes on responsibility for a persons motives when starting up a company...

" A comment like The author is a self-important dilettante. is really nothing more than a pretentious version of "u r a fag."

Well, such a comment is only " a pretension version " if it is not true... and in fact becomes trenchant if, in fact, the respondent IS a self-important dilettante. Which would have bearing on whether or not to engage the person at all... I would not disagree with a person because I thought them a faq, mainly because I don't think a faq is either a bad thing or a thing relevant to many issues I would normally discuss. I would most likely not engage with someone if I thought they were a self-important dilettante, because I DO think that both self-importance and dilettantism is bad. Or, at least, I would say to them, "I think you are a self-important dilettante and I don't have hopes that the conversation will be either enlightening or productive. If we are to engage, you have to rise above your self importance."

Of course, the accusation of dilettantism might be as baseless and rude as that of 'u r a faq' and likewise deserving of scorn. But not always and dismissing it with equal scorn in every instance may be a way of avoiding scrutinizing your own sense of importance and the meaning of dilettantism.

Graham also elides a whole cateqory of 'disagreement' with the implicit assumption that everybody has something worthwhile to say and is therefor deserving of being engaged. "u r a fag" is neither agreement nor disagreement, but simple bad manners. "You are a self-important dilettante" might be a necessary thing to say, especially if the dilettante in question may be in possession of a good brain and otherwise discerning intellect (say on matters not bearing on his/her self-importance...) Intellectual endeavor is approached at with different skill sets, attitudes and skin thickness...

You weren't meant to program computers either...

... nor, I daresay, were any of us meant for great wealth; which, as Graham may yet demonstrate, is just a much a corruptor of character and insight as poverty (if not more so...) Just as large corporations tend to stifle, indeed put a halt to, the generation of ideas, so to does the imposition of large sums of money. If that were not true, then the startups trajectory wouldn't bend towards acquisition... (or, despite Grahams elisions, shall we call it 'surrender'? )

The whole point of evolution is to adapt and mutate and so there is no such thing as 'normal' or 'meant to.' An interesting aspect of evolution is that we get to thumb our noses at the very idea of 'normal' and that we get to shrug off the very idea of 'meant to'.

Paul Graham has argued eloquently and succinctly this very line of reasoning heretofore. It comes, thus, as a shock to read this essay, coming, as it does, from the programmer who advocates bottom up design in opposition to top down mandates. Of course, the point of evolution often contrasts harshly with the point of cynicism: which is simply to justify any childishly satisfying, but nonetheless, foul mood or thought. Or, as Mencken put it, ``The cynic is the man who, when he smells flowers, looks around for a coffin.''

So much of this essay strikes me as the worst form of 'sophistry at a distance' that I scarce can credit Graham as the author. Surely some philosophy undergrad, still pink and slightly damp behind the ears, has stolen into Graham's study and slyly substituted his/her unfinished term paper for some valuable slice of Graham wisdom. Perhaps a wily hacker is having sport of Graham, and us, by subtly crafting an HTML parser that cleverly distorts, ever so slightly, essays fed to it... The world waits, with held breath, for the un-masking of the villain.