Roborooter.com

This post contains a little bit of bragging. My internet setup at work is pretty simple, we have two T1 lines (not counting our voip trunk or our DID lines but that's sort of telephone) going into two different Cisco PIX firewalls and behind those an old Cisco 2600 to do basic routing. One does NATing and port forwarding for our normal internet usage, as well as port forwarding from different IPs for our email and web servers. It's important that the default route to the net not be the same as the email server as when people get viruses that spam everyone people will stop accepting email from your email server. The other provides vpn access to another office, which also only has a T1 Line. And while T1s are slow, it's "enough" bandwidth for our business needs and the other office is in the middle of nowhere and can't do much better. That being said, here in New York City we can do better. A lot better. A T1 usually offers about 1.5Mbit/second for data, I wont cover telephone applications which there are many. That's fine for surfing the net, watching you tube videos, and email. It is slow for 25 people doing all those things, but more importantly it's slow for downloading anything of any size. 1 megabyte for example (about a minute of audio, or one large photo — if you knew that I'm sorry to use the comparisons) takes about 6 seconds. 300 megabytes (for example the size of a decent video clip or a Microsoft or Apple security update) takes about half an hour. 700 megs (say the size of a ubuntu install cd — seriously guys no net install cd? I don't want all your packages.) takes about an hour. You wont see it on this graph as it averages the speed over two hours, but we maxed out our bandwidth quite often. It's mostly my fault, I download a lot. Our network graphs spike all the time and I can say "oh that was me" for most of them. I probably consume more bandwidth then everyone else here put together. It's part of my job (and personality) and because I have to share the connection with 20 other people I can't saturate it for long periods at a time (its rude). At home you probably have about 10Mbit download (700 megs in 10 minutes — but check for yourself) so what slows 20 people down for an hour here would only slow your family or roommate down for 10 minutes at home. Well last week our network graphs automatically adjusted to acomidate a new connection. Have a look at where it says "Maximum" that's 28 times faster then the other graphs maximum. Technically it could read about 60Mbit a second, that's the theoretical limit of our firewall. The Pix501 supports up to 60Mbits firewalled, while the Pix 506E does 100Mbit though its firewall it's busy. What changed was our primary internet connection, we now have a 100Mbit fiber connection from a company called Cogent. They "lit" our building a few years ago but we didn't have the need or $$ to change connections. It's now super cheap (~$700 a month — a bargin compared to the ~$400 for a t1) and has proved to be quite relaible. In a few weeks we're going to move to a Cisco ASA-5505 which will handel firewall, vpn and failover (incase we do loose connection to the internet) drop our remaining t1 line, and steal a few channels off one of the voice T1s for a backup data connection (slow but good enough to keep email flowing). All for less then what we were paying before.

Nice right? Let me put it in perspective. The 700 meg file I can now download in a minute and a half, and when we move to the new hardware it could take 56 seconds. Saving me 59 minutes compared to the origional connection. In actuality we'll probably never hit full speed as most servers wont pump data at 100Mbit/s nor can you guarentee that you'll get routed though the net that fast. There's a noticable speed difference when I pull from california servers compared to new york servers compared to european servers.

My mind is blown. =)

I wanted to email Leo Laporte but he gets too much email. So I figured I'd make it an open email and then twitter about it. Talk about a backhanded way to get someone's attention, but I guess when you're as popular as he is you can only do so much.

Leo,
I was playing with some audio processing tools with a friend and we loaded up your latest twit and saw this. Did you know your intro looked so cool? That's crazy!

On a side note, you can see this for yourself with iPhone app (and I know how you love these) called Spectrograph Lite. You can kind of see the same swigglies that sound so cool.

Enjoy your shows, keep up the good work.

–Francis

(Click the image for a crazy large png)

Really cool looking visualization of the intro to TWIT

Instead of writing up a nice post about the party and showing off my brothers awesome video he cut for the occasion I instead was playing with jQuery. Specificly testing out Ariel Flesler's ScrollTo plugin which I think is pretty cool.

I present to you for your viewing pleasure, my slider test! I couldn't help but use some of Sara's throwie pictures.

And let me tout the video anyway.

Occasionally people come to me and ask about websites.

"Francis, I need help on making website."
"Well.. what are you up to? "
"I have an idea of something I'd like to sell online."

Usually it's cookies, sometimes it's t-shirts, a lot of times it it's services they want to offer. Most of these people have zero programming or html experience. If I told them paypal had an api to allow your site to create invoices and process payments they wouldn't even understand the non technical part of that sentence. When you get into the advanced parts of how the web works, they glaze over. I usually end up telling them about Shopify which can have you up with professional looking storefront with very little time and effort and as you go you can learn how to make it very pretty and you don't have to worry about the perils of doing your own web hosting. Personally I'd rather use that $20 a month for my own server slice and spend a lot more time and money on it, but that's not actually very smart if I wanted to grow a different business that wasn't web hosting.

The real problem they want to solve isn't "I want to know how to do a website." it's "I want to know how to start a business on the web." and the core problem there is usually "I want to start a business." which is usually formed around "I want to have more money". And wanting more money is not a bad reason to start with, but because you like to do something or make something doesn't mean starting a business around it is a smart idea. Most of the time your business will take you away from doing those cool things you like and force you to spend all your time doing something else you wont, running the business.

There's a book I love that will either discourage you or encourage you to start a small business. Either way it will teach you a bit about what you actually have to do to start. I tend to give this book to people who I think it will encourage, as it usually doesn't take a book to discourage people who would be discouraged. It's called The E-Myth Revisited: Why Most Small Businesses Don't Work and What to Do About It by Michael Gerber, and it's like $5 off amazon used so it's worth the money, Its worth the read even if you hate the author's writing style (figure a pound of gold in ten pounds of fluff), and it's worth your time if you ever thing you might want to go into business. The E-Myth being the Entrepreneurial Myth that a technician can take their idea/product and just start a successful business around it. Most people do that without learning how the rest of the business works.

His book of course is not a blueprint of how to start a small business, grow it, operate it, and sell it, it just tells you that you'll need one, and what it might look like. I'm rather new at this so it blew me away. The amount of planning and the possibilities you can come up with to make a business work is breathtaking. Its the kind of thing I love. It's something I'm going to have to write more about. Not to give advice but to share experiences. I can only tell you there's more to it then you probably know, but in sharing experiences you pool what you learn. =)

–Francis

PS (It's been too long since I've been writing, feels good, but I'm also embarrassed at the writing style I presented above. It's a silly feeling.)

I occasionally get into the discussion about why I haven't finished college and why I never bothered to get certified in anything. I like to believe past performance is a good indicator for future performance. I have a decent resume with a few good jobs that I've put a few years in on. I try all sorts of things on my own all the time. I want to go for jobs that allow me to learn and keep expanding my repertoire. I worry that I don't have the skills for those jobs but that's sort of the point. I don't think any piece of paper could show any of that better then my resume, certainly not a degree or certification.

Anyway (and yes I still think about finishing school mom and dad), I read this interesting article called After Credentials by Paul Grahm of YCombinator (an interesting VC firm specializing in startups). It's a nice short read about this history of why he thinks they came about and where they're going. Let me excerpt it for you. (I skipped everything, these are very non related paragraphs. Also I skipped the part about the dawn of the yuppie and how it's not odd to see a 25 year old with money anymore.)

Before credentials, government positions were obtained mainly by family influence, if not outright bribery. It was a great step forward to judge people by their performance on a test. But by no means a perfect solution. When you judge people that way, you tend to get cram schools—which they did in Ming China and nineteenth century England just as much as in present day South Korea.

The obvious way to solve the problem is to make credentials better. If the tests a society uses are currently hackable, we can study the way people beat them and try to plug the holes. You can use the cram schools to show you where most of the holes are. They also tell you when you're succeeding in fixing them: when cram schools become less popular.

This doesn't work in small companies. Even if your colleagues were impressed by your credentials, they'd soon be parted from you if your performance didn't match, because the company would go out of business and the people would be dispersed.

Credentials are a step beyond bribery and influence. But they're not the final step. There's an even better way to block the transmission of power between generations: to encourage the trend toward an economy made of more, smaller units. Then you can measure what credentials merely predict.

Go read the article and comment here. =)

–Francis