Archive for posts tagged ‘web’

Amazon S3 and CloudFront Response Time

Kevin Day, April 8th, 2010

If you’re evaluating whether to serve static files from Amazon S3 or their CDN, CloudFront, here is some data I’ve collected for each of their response times.  The graph below is the result of two tests I set up in Binary Canary:

s3_vs_cloudfront

It’s a bit hard to read, but it says that CloudFront’s average response time is 164 ms, and S3’s response time is 338 ms.  The test pages I set up were small 60 byte HTML files.

There are technical factors that need to be considered when choosing one or the other, but it looks like you can shave off about 160 ms by choosing CloudFront instead of S3.

Contact Management Software

Kevin Day, November 22nd, 2009

There’s a lot of contact management software out there.  I don’t know if it’s just me, but none of them seem to fit the way that I work.  Just managing the contacts for my fantasy football website quickly becomes overwhelming.

That’s why I’m developing Launch Pulse, a new kind of contact management application.  It’s going to have about 1/20th of the features of all the other programs out there.  And yes, I am aware of Highrise by 37Signals.  Fewer features than even that.  I’m not sure yet if it anyone else will feel the same about it, but I like using it already in it’s alpha state.

How is it different? It won’t sync with your Outlook or Gmail contacts.   It isn’t tied to your email system at all.  But it is flexible.  I think the initial target audience will be people using Customer Development because it requires talking to a lot of people and managing responses and feedback.  However, I think it could later be expanded to manage contacts for weddings, job searches, and anything else.

It’s not available yet, but you can leave your email address on the website (no spam) to get notified first when it enters public beta.

Micro-innovations that make me happy

Kevin Day, March 12th, 2009

Two things came out this week that will instantly make my life easier and cheaper:

Da Button Factory: Instantly create shiny buttons online.  I had no idea how bad I was at making buttons until I spent 5 minutes on their site and saw what I could do.  Also, within the past two days they’ve already made several improvements to the site.

Amazon Reserved Instances: The same Amazon server that I’ve been using is now going to cost 33% less with their new payment plan.  A virtual dedicated server (not including bandwidth) now only costs $48/month.  Awesome.

Great domain name checker

Kevin Day, January 15th, 2009

In the past I’ve used pcnames.com for searching domain names.  They have a lot of good tools there and I’ll probably continue to use them.

However, I just found bustaname.com and it’s pretty awesome.  It has a synonym finder, saved searches, grouped words, prefixes and suffixes. It’ll probably be the first site I go to from now on.

As I was using this, I thought that another extension would be to have the option to do two more searches:

1) Regular Google search for those terms, since there could be conflicts with major companies/organizations

2) U.S. Trademark search to avoid nasty IP issues.

#1 would be easy, not sure about #2 though.  Compared to getting a name I like, those aren’t nearly as hard though.

(FYI: Referral code in the link above)

New Job

Kevin Day, December 27th, 2008

I’ve recently started work as a web developer for iGuiders, a guided search company.  I’m excited because it’s a small, focused team working on a product that has a lot of promise.  I’ll post more details after we launch.

The Power of Labeling Images

Kevin Day, October 6th, 2008

In my Google Analytics account, I noticed a few visitors arriving via Google Image searches. It turns out that this picture got a lot of Google juice and is the first image search result for the term “smiley pumpkin“.

What’s surprising is that that term doesn’t appear anywhere in the page or this site except for within the image tag:

<img src="http://codeswimming.com/images/smiley_pumpkin.jpg"
 title="Smiley Pumpkin" alt="Smiley Pumpkin" width="500" height="375"/>

This blog only has a page rank of 1, yet by including the same search term in the file name, image title, and alt text, it got to the top of Google, albeit for a low-volume search term. There also don’t seem to be any links to that image from any other site that I can find.

This only has limited SEO use, however, because that post and my blog are not in the general web search results for the same term. Google seems to keep the two sets of results completely separate.

Thank God! Amazon persistent storage is here

Kevin Day, August 21st, 2008

I just stayed up late last night working on backup scripts for my EC2 instances, so it was great to wake up this morning to see Amazon’s latest update. With Amazon’s Elastic Block Store (EBS), ECS users can now mount persistent volumes directly to the file system. That should translate into more sleep for me.

I don’t know how many I/O requests I make in a month though, and that seems to be how they’re charging for this service. I’ll have to test it out first before I do anything serious with it.

Firefox Caching Bug

Kevin Day, August 17th, 2008

For a while I thought I had a bug with my Ajax fantasy football mock draft program, but it turns out the bug was in Firefox 3 instead.

The bug is that Firefox 3 doesn’t handle caching properly.  If you don’t want it to cache a page you have to specify your headers in a specific order, otherwise it will cache it even if the headers correctly specify not to.  More info here and here.

This affected me because I use a periodic Ajax GET request to the same url in a live draft application.  I reply with no-cache headers so that the GET request polls the server every time.  When Firefox 3 pulled the request from the cache instead, it would miss a lot of updates and made the application appear to freeze up.

To work around the bug, I’m just appending a timestamp to the end of every request.  I feel that’s better than using FF3’s specific ordering of header info because I don’t want to have to worry that they may change their magical ordering.  It’s not much work to get around the bug, but it was difficult to pinpoint in the first place.

DIY Startup School Cleveland-style

Kevin Day, May 22nd, 2008

Although the official Startup School was last month at Stanford, there was enough startup activity in Clevleand last week to constitute it’s own mini startup school.

Thursday was Cleveland Startup Drinks, which was a great chance to talk with other startup founders.

Saturday was the SCORE business workshop that was a half-day of informative talks by an accountant, a lawyer, and a bank president about how to develop a business plan, incorporate a business, and finance a business. This was the first time I had heard of SCORE, and it looks like a valuable resource for new businesses. They are a large group of retired business executives that offer free counseling. What could be better than that?

Lastly, on Sunday morning Craig Newmark spoke at the CWRU commencement ceremony. It wasn’t quite as inspriational as the real startup school talks, but it was a taste of Silicon Valley nonetheless.

Negligent web surfer

Kevin Day, March 1st, 2008

When I was reading football news today, I discovered that the NFL has a Chinese website, nflchina.com. It’s mostly news articles that are translated into Chinese.

I used Google to translate the text to English, unsure of the results. I think it did a decent job. Here’s the Cleveland Browns page:

However, when I was browsing the site I got this error message at the top of a page:

You did not examine the competence of this website. Please sign.

I guess I’ll have to do a more thorough evaluation next time.

Also, I was disappointed that Google wouldn’t let me translate an already translated page. I wanted to see how the translations of translations would turn out.