lethain

Thus far they don't delete tweets from the search results even after they have been deleted within twitter itself, so I suspect they are append only as you suggest.

The inverted pyramid is effective at conveying information quickly, but generally essays are not breaking news, but instead rehashing or adding a new perspective to something old. In the later scenario, I don't think the inverted pyramid is as effective.

He's basing that statement on a study performed on a number of runners which tried to identify factors that would make someone likely to have an injury over a year, and the only one that was statistically significant was price of running shoe.

Certainly one study doesn't provide a definitive answer, but decrying the article because he is inconsistent with common sense, which is he is explicitly saying is wrong, seems a bit unfair.

I don't think we can argue that it isn't natural. Before the last century or two it was rather natural, among other species it is quite common for genders to have specialized roles as well, and there are certain roles which will always be imbalanced (childbearing, for one).

As I understand it, in the end the objection to gender imbalance is that it constitutes an unfairness to either the minority (executives, programmer) or majority (nurses + school teaching during the period their wage/benefits were still severely depressed by gender imbalance).

I don't think we can argue that unfairness is a threat to the system or to groups, it is universal. Rather, I think we just prefer not to realize it's there.

You've mistaken me for my argument a bit there. The right's fiscal philosophy is based on the idea that people who earn money deserve to keep it. People who don't earn money shouldn't be re-appropriated money from those who do. In the long term that is achieved by having a smaller, less expensive government.

It is not the case that the government must raise taxes when they exhaust their resources, the other option--the desirable outcome from small government advocates--is that they slash services and oversight to an extent that is politically impossible under other circumstances.

Politics is not motivated purely by reason, but also by the different parties are motivated by philosophical ideals.

When done tactically, government spending can lead to smaller government. The long term goal for much of the right's spending agenda is to create an environment where the government is forced to become smaller via debt fueled program slashing. Defense spending is particularly good for this, because there is abundant rhetoric for attacking someone who reduces (i.e. "cuts") defense funding.

django-compress does this very well (integrates with YUI Compressor, and several other compression filters). I use it on all my django sites.

The problem with an engineering student telling a liberal arts student that their degree isn't valuable, is that the engineering student's opinion is very rarely grounded in experience. Although one group happens to have the "right" opinion, that's all it is until reality vindicates it.

Also interesting, one of the links on the site leads to another discussion which claims that there is a correlation between higher intelligence and lower wages for those whose highest degree is a bachelors. The study partially attributes this to the bad advice that more intelligent students tend to get from their parents ("study what you love", as opposed to "study something practical").

(Referenced link is http://www.halfsigma.com/2006/07/higher_intellig.html )

Stallman surely realizes that minifying JavaScript is a standard process, but it doesn't seem particularly relevant to his point. First, it not straightforward to determine the licensing for many scripts, and second it does not provide access to an unobfuscated copy of the source. In no way does he ask for JavaScript to be served without minification, but rather he asks that minified JS contain a pointer to a non-minified copy. Doing so doesn't create an additional cost for the site owner, beyond possibly requiring an extra step in a build script.

I think you are pushing two distinct movements into the same box: the free software movement championed by Stallman, and the "get users first, monetize later" euphoria of the "Web 2.0" movement. The first group is pursuing a moral agenda; the second believes that it is a legitimate route to wealth. Much of what you are blaming Stallman for seems better directed at the second movement or at open source developers (who are often a very different breed than Stallman, and who--in my experience--tend to hate the GPL as an unnecessarily restrictive license).

I also find your example of iPhone and Facebook applications as somewhat problematic. These platforms are ghettos where developers are dependent on the capricious rules and regulations provided by the owners. This is hard to ignore when you see traffic to applications dropping by a reported 50% after the latest redesign for Facebook, and also when independent apps like Podcaster (err, whatever the name was) barred from the AppStore because it competed with functionality that Apple's own apps provided. Opportunity abounds in both, but there are many many losers in both as well. The real winner in both these situations is Apple and Facebook. Asking forgiveness from Marx, when possible it is always best to be the landlord.

My second to last point in this unfortunately disconnected response is that neither Stallman nor Web 2.0 have ever been the "guard". Both are and have always been fringe movements. The "guard" has always been large companies selling software to other large companies and making tremendous sums of money while doing so.

Finally, people do get paid for open source contributions. It isn't straight forward, and the correlation between time spent and money returned is not predictable in the same way it is for an hourly consultant gig, but there is always money there for people who grasp it.

Do you want this individual as a co-founder?, because that is essentially the role they would be stepping in as. At minimum you would need to write a solid vesting plan of some kind, to prevent the individual from quitting in a month and taking the equity with them.

That said, I would be somewhat suspicious of the situation where someone is willing to work for only equity. Are they already independently wealthy? Are they having trouble finding paying employment?

Yep. They prorate charges, so firing up an extra slice for a couple of days will only cost a few dollars.

I've been running my sites (Django-powered) on Slicehost for about two years and have been very pleased overall.

A few observations:

1. Running Apache with both mod_php and mod_passenger (assuming that would be your deployment method for Rails app) on a 256 meg slice may be tight if your traffic levels pick up. (Based on experience with mod_wsgi, mod_python and mod_php, I haven't explicitly used mod_passenger and am unaware of its memory/interpreter model, but think it is similar to mod_wsgi.) But it will be enough to manage your existing blogs and additional PHP applications as long as the load doesn't explode.

2. You don't need any linux administration chops to get started, but every year or so I've learned enough additional practices that I grab another slice, setup the new slice from scratch, transfer my webapps/sites to the new slice, and power down the old one. There are just a number of small best-practices that--when avoided--add up to a hard to manage setup over time, especially when managing multiple sites within multiple servers. You'll probably want to do the same.

3. Slicehost has great resources, http://articles.slicehost.com/ , which should cover most configuration/setup details.

4. Switching over to my own slice and doing administration myself has been a very valuable experience for me, and I really do recommend it.

For what it's worth, this is a side project which has given me a lot of perspective on my professional work as well, but is certainly not a "start up" or otherwise intended to directly create revenue.

Although that sounds correct, I suspect it's difficult to argue the distinction between an image download and an image cached by the browser to people who are extremely untechnical.

Out of curiosity, why is it that you're looking for an internship? Is your previous experience not related to development? Or is it simply the timing issue, that you don't feel you could find a non-internship opportunity given your narrow window of 2-3 months to work?

I didn't get to explore your service too deeply because the UI really needs work. For example, the "more services" button just unleashes them onto the screen, forcing me to scroll the screen to view them.

Also, switching between different searches doesn't react consistently, which creates a very bad feeling for the user as they can't reasonably predict what their actions will do.

There are a number of areas where Findjango--or vertical search engines in general--can outperform a mainstream search engine. It's certainly a work in progress.

1. "Perfect answers" that relate to the topic at hand. For example, Findjango has an index of Django's documentation and lists related documentation at the top (above search results). For a large percentage of searches, those links will answer the search immediately (also, the links deeplink to the correct location on the documentation page, as opposed to simply dropping you at the page's beginning).

2. Because of a narrow scope, it is possible for a vertical search to be more current.

3. Because of the narrow scope, it is possible to apply custom relevancy algorithms suited to the specific niche which may outperform generic relevancy algorithms. Findjango has begun traveling this route, but is clearly not there yet.

4. The ability to collect and display targeted metadata. Although I'm not taking advantage of it yet, Findjango knows the number of followers and last commit date for each GitHub repository, and can display them in-line in the SERP. For a targeted audience this can be worthwhile to add, whereas for a main-stream search it would represented a negligible usecase.

5. I believe that vertical search engines can be monetized much more effectively--and while doing so create more added value for the user--than a generic search engine. Certainly I have no numbers to back this up as of yet.

In the end, my argument is that a vertical search can outperform a general search engine within its vertical. It is certainly not my argument that it will inherently outperform, and certainly Findjango lacks the breadth of content to compete thus far.

This is a long-term project, certainly not an overnight road to success.

It's important to keep in mind that--in the absence of downvotes--scores only mean that content is relatively more relevant than other content surrounding it, not that it is absolutely relevant to our audience. The current voting system encourages diversity, and it's diversity we get.

Josh Gourneau setup a MyCroft plugin which makes it pretty trivial to use Findjango 'in browser' for IE and FF, but not sure what the options are for WebKit or Opera.

I agree that exposure is a huge issue there, as we're not likely to sign and distribution deals in the near future.

Although few sites buck the tradition to have the www subdomain be handled identically to primary domain, it is not a requirement to do so. I don't have examples flowing out of my pocket, but it's just a DNS configuration away from reality.

You could also phrase this as the distinction between being "correct" and being "right". The current solution is "correct", although the proposed solution will typically be "right".

You're right that most of the time www.example.com and example.com are equivalent. But they aren't necessarily equivalent, which is why assuming they are identical is problematic.

Your suggested modification will be correct 99% of the time, but will be incapable of handling the other 1%. The current system is correct 50% of the time, and can correctly handle the 1% that the proposed modification cannot.

Philosophically the current solution seems to suggest it is better to be less correct on average but correctly handle edge-cases. As another consideration, the userbase can correct mistakes in the current algorithm, whereas they cannot correct mistakes in your proposed algorithm.

I think this frustration comes from feeling that you're "playing by the rules", and believe that those who don't follow the rules should be punished. (Or at minimum, those who do play by the rules should be rewarded in such a way that they outperform the cheaters.)

This line of thought has only made me unhappy, and is deviously self-serving at always snatching a moral victory out of defeat. Don't worry about those who take the bus while you're out jogging.

There are a whole mess of questions wrapped up in this project, some of the interesting ones are:

1. Can a targeted search provide a better experience than G/Y/M?

2. Even if that search is a better experience, would anyone use it?

3. Managing relevancy on results from unreliable sources.

I think, depending on my integrating more resources and greatly improving the handling of mediocre resources (i.e. the current results of the solr based search for my blog's content, which are often totally irrelevant...), search verticals will be an interesting trend to watch.

I imagine it's less than you're suggesting, because they're not serving the files themselves, so their bandwidth would be relatively low compared to a photo or video sharing site.

I agree. People will read good content on a poorly designed blog. And if they like your content but hate your design, they--in my sad experience--make it very clear that they'd like you to improve your design. A great design for a blog is kind of like great office space for a startup, it's nice to have but is a misuse of resources if done prematurely.

I think he would include that indirect revenue as "a way to make money off of your blog."

Although Python has been leaping down the slippery slope as well, I think calling Ruby's syntax simple is a bit of a stretch. Ruby has many--intentionally--more ways to do everything, and some of those more ways have small gotchas built in.

Although Google probably won't help you here, you can help yourself. If your campus traffic shares an ip or ip range, you can just tweak your site to not show the ads for those sites, which should be pretty straight forward.

This is what we call optimism. Just as having both risk-loving and risk-averse people creates a more stable species, so does having optimists and pessimists.

Its a perfectly reasonable thing to believe--just like its opposite--and depends solely on how you're measuring. All in all, I think it depends on the reduce function. If it was reduce(lambda a,b: a+b,[ x.happiness for x in humanity ]), then maybe there is the most happy the world has ever seen. If you took that number and divided it by the population, than maybe the average happy per person has gotten lower. And so on and so on.

I think its definitely possible to believe it.