stuki

Since no one else seems to be doing so, let me put in a plug for MSSQL. If you're using the Microsoft stack, it ought to be the default choice. The integration with the rest of the environment (LINQ, db process LCR, Visual Studio integration, async transactional queuing, one heck of a molap engine, a great etl tool and more) is just too good to give up. And it's a darned nice database to boot, with 2008 on the way being even nicer. Also, Microsoft has publicly assured customers they will continue to license per socket, not per core. While still more expensive than the free db's (duh), at least you're not stuck handing over all future Moore's law gains to Larry Ellison.

I'm becoming more and more convinced that one of the keys to getting good software is to have it developed by a software company, not by the technical department of a non software company, or even by a multifaceted 'management consulting' company. Unless developing software is a company's sole concern, hierarchies, rituals and reward structures certain to drive away the best developers, almost inevitably take root.

Dayum, no wonder some of these venture guys want to see a 'business guy' on board. I somehow doubt I could get a check out of this one if he asked about my Revenue Recognition Policy, and I answered: Uhh! Dude! Wanna check out some really cool hofs I wrote last night? :)

On a more serious note, if one actually plans on making a living off of a software startup, without betting on simply getting acquired at a crazy valuation; at some point I guess figuring out at least roughly where the money will come from does become a bit of a necessity. Even if that means accepting revenue, at least partially, being a function of number of sales guys.:(

That is definitely true, and probably a big differentiator between Silicon Valley and Wall Street. In Silicon Valley, there are those who are rich 'accidentally', simply as a side effect of doing something they really wanted to do. You see the same in the entertainment business, as well as in 'leisure' industries, like mountain bike and surf board building.

Once you go beyond these 'legends', to the, I believe, much larger group of people described in the article, this no longer seems to hold, though. The "legends'" possibly inadvertent, yet widely publicized, financial success, seems to have drawn to The Valley many more that ARE in it for the money, and these are the ones I doubt will ever be content.

Seems to me, people are either of the kind that's happy as long as they have a cold beer and someone to share it with, or the kind that will always be sufficiently unhappy with what they have to bust their ass in pursuit of more.

I suspect Silicon Valley, like New York, is home to an over abundance of the latter kind. While at the same time, those who can't understand continuing to work like that despite having millions, are exactly those that would never have bothered putting in the effort to get millions in the first place.

It could very well be that, for most people working 'inside the bubble', this one may, in retrospect, have ended very similarly to the last one. A lot of people essentially spent a bunch of years working for chump, in widely shared anticipation of riches that never materialized.

At the same time, since not nearly as many outsiders are betting their life savings on the 'bubblers', the economy wide impact of even the most devastating sector wipeout, won't be nearly as bad as dot bomb.

In that respect, how you'll come to view it will depend largely on where you stand.

It could also be that what we are seeing is the leading edge of PG's predicted trend for more young people to enter working life by doing a startup, rather than going to work for someone. This would probably in and of itself lead to an economic landscape looking a bit 'bubbly' by historical standards.

I guess only time will tell.

For the perhaps less hardcore, and less Blub averse, F# is a caml implementation/extension? on .net. Less/no need to write homegrown, production quality c/c++ interfaces, since it integrates with the rest of .net. It will compile some/most ocaml code without too much change. I can't help but think using it, or at least being aware of it, can benefit some software startups.

That was my point exactly. I would think anyone setting up a wide ranging electronic payment system must need some sort of funding, banking deals etc. Being US based, getting that with a prominent link for Online Casinos etc. to sign up as merchants on the front page, seems rather unlikely. Hence, unless they refocus, the whole startup seems a bit unserious. I wish them the best of luck, though; hoping that does not make me guilty by association:).

After the whole NETeller debacle, I would be wary of association with anyone providing service to online betting shops. http://www.wnbc.com/news/13656319/detail.html?subid=10101421

It would be interesting to see how MS's kernel differs between client OS's and their servers. They have huge user camps both on the desktop and in enterprises, both of which are major, strategically important profit centers.

The GPL pretty much ensure a server centric development model on Linux, though. Most open source developers and supporters aren't quite so 'open' with their own companies' source code, thus preferring to 'distribute' their apps as web apps, requiring no source disclosure.

Theoretically, as search engines and advertisers both get more and more sophisticated, the organic and paid for links should converge. The most 'relevant' results for (search term, searcher history, other relevant data) indicating someone looking to buy something will be the sellers most likely to convert the given searcher. At the same time, advertisers would be unwilling to pay for clicks from searches indicating unlikely buyers. If matching ever get that sophisticated, some altruistic sounding blurb about keeping commercial links off the organic section probably won't be far behind, however.

Sounds like a nobleman griping about commoners' ability to crash the party with cash. Venture financing would be a strange animal indeed if there was one true best way to structure every deal for every venture. If there is a big demand for this, the big boys refuse to serve it, and being funded by the big boys truly increases success probability substantially, some clever Wall Street'er could find a way to buy part of, say a Sequoia funded, founders upside; or at least something somewhat aligned with it. Even if it means working around some legal headaches. As an aside, these founder liquidity events could be just the ticket for those much maligned 'idea guys':).

Wow! And here I was thinking $3 billion for Doubleclick was rich!

Whatever the merits of the acquisition on its own, deals like this is bound to send a strong signal to entrepreneurs and investors that the 'new' MS sometimes will buy rather than crush. Since they're sitting on the biggest cashpile this side of China, that is a sweet thing indeed!

Heck no. The runtime is state of the art. So are many/most of the important libraries. Until another environment available ex Windows get close to that, java is solid. Generally speaking, languages with 'nice' features are a dime a dozen. It's the runtimes and ecosystems that are hard to duplicate.

'Worlds most X' programming language, for programming what? The ideal language in which to express something kind of depends on what you are trying to express, doesn't it? ( maybe unless you are some kind of meta lisper ( or perl6'er :) ) dreaming of a language that's a strict superset of all possible others )

pg, whatever you read into this rant, please don't let it make you go soft and pull your punches in essays, posts, talks or interviews. Most of your audience prefer your straight talk to mealy mouthed, watered down, PC blandspeak, even when you're wrong. And if destitution necessitated scurvy (out of all things......) turns out to be a huge problem amongst your founders, I suspect there are investors out there willing to come to the rescue, even at twice YC's valuations. No questions asked :).

PS: for those who find the IQ test comment arrogant; Paul Graham is just showing off his mastery of Cambridgeian lingo. Back there, saying someone has low IQ is short for saying they disagree with you. Nothing more, nothing less, and absolutely nothing to be all up in arms about.

Hopefully, this time the dispute will go all the way to the supreme court, and we will get a ruling that makes the whole IP environment more predictable. At least for startups and smaller players without huge legal resources, as well as for those who would like to invest in them, the current uncertainty is worse than any probable ruling.

When Your immigration lawyer tells You You have had your green card long enough for it to 'vest' ( 18 months ) :)

....And for the good of the poor schlubs who ask for nothing more than decent quality video and well performing, featurefull apps delivered to their computer or phone, I for one hope whoever can deliver an infrastructure capable of doing that thumps whoever can't. Standards are a nice way of encoding minimal required commonalities for a class of something, but should never be used to prevent anyone from going above and beyond.

I also suspect MSFT's ability to cram anything down the throat of subjects opposed is highly overrated. I personally bought my first windows machine a few months ago, having been (still am) a mac, sometime linux, user for years. If Silverlight (or flex/flash) gains huge penetration, it is likely because people prefer having it to not, not because of any 'cramming' by MS/Adobe.

In my admittedly old fashioned view, the killer app for phones is talk, and the competition is too entrenched :).

Dudes, unless the silverlight demos was pure smoke and mirrors, and all involved blatant liars, this is pretty cool stuff. Especially coming from a bunch of dead guys :). This is something I have been hoping for from MS for a decade: website like ease of app delivery combined with a non stone age client programming model. MS is the only one with the size and reach to have any kind of shot at pulling it off. Nice to see they are finally taking their 'monopoly responsibilities' seriously. Supposed monopoly or not, size really does matter when it comes to providing us little guys with assurances that decent codecs and runtimes are present on our target market's computers ( and phones, gaming consoles etc. ). How any developers not in direct competition with MS can not like being able to do rich clients in .net + assorted tools instead of the current mix of hacked together technologies is beyond me. This doesn't mean every app has to utilize a rich client, just that for the ones that do, we can build them in a modern development environment and expect modern runtime performance. Now..... arc.net? Orders of magnitude faster than JS...

Ram disks (like www.texmemsys.com) for the log device are a way of speeding up traditional rdbms'. They cost a bit, but You get to keep all the 'cool' database features while minimizing the sync log write bottleneck. Such architectures are also easier to explain to auditors if Your app is of a financial nature. I am not affiliated with this or any other storage vendor. --Stuki