alex.moskalyuk

SmugMug is the poster child for Amazon’s Simple Storage Service, known among us simple folks as Amazon S3. Even though some other startups rely on S3 for their storage (I noticed MyBlogLog pulling their stats and data off S3 recently), SmugMug has been pretty vocal about their S3 savings.

So what about the recent outages that S3 is becoming notorious for? Deal breaker, or cost of doing business? Don MacAskill tells the story:

S3 is a new service and yet remarkably reliable. Since April 2006, they’ve been more reliable than our own internal systems, which I consider to be quite reliable. Nothing’s perfect, but they’re doing quite well so far for a brand-new service. Oh, and their services has also saved our butts a few times. I’ll try to write those up in the near future, too.

Over at XAPRB blog there’s a great tutorial and discussion on using MySQL FEDERATED storage engine, which abstracts the data access layer among the different servers and allows you to access a table on a remote server as if you were using a local server with no explicit outbound connection to the remote MySQL server. There’s a pretty comprehensive article on using FEDERATED storage engine over on O’Reilly Network. There are quite a few findings about the FEDERATED tables over ay XARPB that will give you a better understanding of when the engine should be used and when it should not:

Rows retrieved from the remote storage don’t seem to be cached even for an instant. For example, if you issue a join against a remote table where the local table contains repeated data, the matching rows will be fetched over and over again from the remote table. One consequence is that if something updates the remote table while this is happening, you will see an inconsistent view of it, even within a single query.

From now and till April 30th of this year HSBC Direct will offer 6.00% APY on new money. If you already have money with HSBC Direct, the old APY of 5.05% will apply to whatever was on the account by January 26 end of day, and all the new deposits will get the preferential 6.00% APY. I am a happy HSBC Direct customer myself, and one of the advantages of having an Online Savings account with them is ability to link up to 8 external accounts.

PHPBuilder runs a chapter from an APress book on PEAR. The chapter is dedicated to using PEAR authentication modules. Pretty much any site you build nowadays allows the users to register, choose a password, validate a password, send a password out when it’s forgotten, etc.

You use the Auth package to authenticate users in your site. Out of the box, it supports many different ways of authenticating users, including storage in a database, in files, or even by using SOAP calls. You can even write a custom container object that allows you to write your own method to authenticate users.

Donald McArthur over at NewsForge creates a generic reusable PHP calendar template that could be used on any site requiring inputting dates:

My design goals were to create a PHP page that would take as input a querystring value in the form of a Unix epoch number that would represent the beginning moment of a particular date. (I chose the Unix epoch number, which represents the number of seconds that have transpired since the start of January 1, 1970, as that was the data my database SQL statement used as a SELECT criteria.) The script would determine the month and year of that value, and create an array holding a Unix epoch number for the beginning moment for each day in that month. The script would then output HTML to display a calendar, with each date a hyperlink back to the original PHP page, with the associated querystring value for that date.

There’s a heated discussion over at ThreadWatch regarding Google AdWords going live on MySpace as part of the search network syndication. It seems that a lot of advertisers are not exactly happy about this, as you cannot opt out of MySpace search ads on a domain-by-domain basis. You can certainly tell Google to stop showing your ads on search network and stick to google.com searches, but that also kills nicely converting AOL Search.

I have never had a MySpace account, or any reason to browse profiles there, but one of my sites that’s done fairly well on Google recently started getting a healthy dose of traffic from MySpace. The referring link in the logs is http://sads.myspace.com//Modules/Search/Pages/Search.aspx? with the query following. That page is generated whenever a MySpace user conducts a top-level search box query, and to be redirected to my site, I assume, they have to click on the Web search as the default limits you to MySpace contents.

What’s peculiar about those searches is that 90% of them are MySpace-related searches, and none are those generic Web searches that drive traffic from Google. I don’t run a MySpace-specific site, but the user discussion boards have enough content in them to come up in Google for searches like “myspace layouts”, “myspace codes” and other paraphernalia. Only by a matter of coincidence pages from my site rank on Google on those terms (usually because someone points to myspace.com in their signature, and there’s a word “layout” somewhere else on the page), so being ranked in 40+ range I never get any real traffic for those.

Except when it comes to searches generated from sads.myspace.com. Not only those are almost exclusively MySpace-related, the users apparently go through the trouble of navigating 4-5 pages deep into the result set to click on a site that might vaguely have what they’re looking for.

From this personal anecdotal example (feel free to chime in if you get traffic from sads.myspace.com) I think very few of MySpace users view that top-level search as launchpad for their Web searches. Most of them use it for internal site search, only clicking on the Web tab, when the initial search produces nothing worthwhile. Which is good news for advertisers, as the possibility for erroneous or irrelevant ad click is reduced, unless you’re specifically targeting MySpace audience.

I am halfway through the 30-day free trial of BlockBuster Total Access, and by now I am quite ready to cancel out of Netflix, even with forthcoming Internet streaming feature. There’s a BlockBuster participating in Total Access plan right across from Facebook worldwide corporate headquarters, and living within a walking distance contributes to the bias towards Blockbuster. Your mileage may vary.

Every movie I get via mail I return to local Blockbuster and get a free in-store rental, which usually allows me to get the newest releases right away. Moreover, the return gets recorded in Blobkbuster database, the next movie is queued for being sent out the next morning, and effectively I am getting double the rentals I was getting with Netflix.

This is especially relevant for the weekend, as USPS doesn’t deliver on Sundays, so frequently sending out Netflix envelopes on Friday or Saturday would not get your replacement DVDs in till Monday-Tuesday. With Blockbuster I just bring Friday/Saturday mailings to the local store.

A coupon for a free rental or game doesn’t hurt either. Granted, it’s once a month thing, but it allows me to test drive a Wii game before committing to buying one.

Previously it seemed that Blockbuster selection was not as wide as Netflix’s one, so if one’s after the long tail of movies, Netflix was a definite choice. Now I browse through the New Releases pages of both sites, and have not noticed any significant differences between two online stores. Basically, there’re just so many suppliers in the wholesale DVD world, and looks like both Netflix and Blockbuster cover the same selection. Neither one, as far as I know, has any exclusive deals, so in the long run both will have to strive to avoid commoditization.

Microsoft rebranded Atlas as ASP.NET AJAX and made the reusable components available as a public release. I vaguely skimmed over the news headlines, since Active Server Pages dot net Asyncronous JavaScript and XML did not hold any promise of being too useful on the LAMP stack, but now CodePlex released some libraries to allow PHP to interact with Microsoft’s reusable components:

require_once '../../dist/MSAjaxService.php';
class HelloService extends MSAjaxService
{
function SayHello($name)
{
return "Hello, " . $name . "!";
}
}
$h = new HelloService();
$h->ProcessRequest();


I tried a different approach to reading a programming book recently, and it seemed to have worked out better. I started reading from the last chapter and then moving back. If it’s a typical book introducing new technology or language, most likely the last few chapters present some advanced concepts about the language or technology, and some books include some heavy-weight practical projects that are applicable for the real world.

What do you usually get when you read a book from the beginning to the end? A whole bunch of introductions to such concepts as variables, functions, loops, conditional statements and benefits of object-oriented or whatever approaches the author advocates. The information value of those chapters is very minimal, unless you have not read a single programming book previously, and this is indeed your first language or technology.

When you start at the end, you’re usually presented with

1) a good description of the best usage scenarios for the technology
2) a sample of how well-written code looks like in this language
3) all the shortcuts and more or less real-world code used, no one is spelling out if-else statement with extraneous comments anymore
4) a good feeling of whether you generally “dig” what’s going on

Try it out. Pick up some random book (Python, C#.NET, Rails, etc.) and head off to the last chapter. I think that you’ll be less bored overall and will get a better dive-in introduction than following the first few chapters.

Today Associated Press ran an article on how Cisco Systems benefited from online video explosion and apparently is either considering new product lines or vastly improving the existing ones to accommodate the YouTube phenomenon:

Charlie Giancarlo, Cisco senior vice president and chief development officer, said the push also is forcing Cisco to change the way its products are built to deliver high-quality video, which puts more strain on the network than voice or data transmissions.

What’s interesting here is that the flagship online video site - YouTube, has never made a dime of profit, and depending on the server bills and advertiser saturation, might never reach green. Same goes for Google Video and other large scale video projects, whose bandwidth-to-ads margins are, one could imagine, pretty slim. At the same time there are these “hidden” players of the industry, who benefit tremendously, since their stuff sells and sells fast. Cisco is one of them, another hidden flagship of the online video industry is Akamai, whose 12-month share price going from $21.88 to $54.64 looks pretty nice to any investor who got in at the right time.

So if you judge online video industry by the money that YouTube and Google Video made, it’s really insignificant. If you judge by the sales that Cisco and Akamai made, all of a sudden it’s multi-billion-dollar industry.

Same goes for some other industries like e-commerce and FedEx. If you track e-commerce trends by the sales and margins of Amazon.com, Staples.com or other players, you might think that the margins are tough and price competition is enormous, therefore being in e-commerce is more of a curse than a solid business model. But both successful e-commerce shops and unsuccessful ones ultimately turn to their FedEx guy for shipping, and a 5-year curve of FDX follows US e-commerce volume closer than that of AMZN or other significant pure-e-commerce players.