Sunday, September 25, 2011
Web Hosting Deal of the Year
Dreamhost.com seems like an honest and reliable shared web host. They normally cost more than some of the other hosts, but today I visited their site and they're offering a $110 discount on the first year of hosting. So you can get a domain and a full year of shared web hosting for $9.24, which is about the cost of a domain alone. That's the best deal I've ever seen, though the second year will be at the regular price. An alternative option is Dreamhost Apps, which offers free webhosting for Wordpress and a couple other popular scripts. However, it seemed a bit buggy when I tried it out, and its definitely not like a full-fledged webhosting account.
Thursday, September 22, 2011
Google's Idea, the Quality of Search & Quantum Mechanics
Google's been in trouble lately for being too successful, so I thought it would be worth examining how Google improved search and what quality issues have arisen.
Historical background
In the 1990’s the Internet changed the world by providing a way for every type of information to be accessed from anywhere. Yet, this plethora of information would have been impossible to use without technology capable of sifting through it to find what the user needed. Many search engines tried different techniques to rank search results, but none of them did a satisfactory job.
In 1998, Google entered the scene. While Sergey Brin and Larry Page were graduate students at Stanford, they developed a new way to rank search results and created Google.com. However, Google soon became a distraction from their important work pursing a PhD, so they tried selling it for $1 million, but no one wanted to buy it. Sadly, they were forced to drop out of school to focus on their company. By 2007, their company was worth more than $200 billion, but they still had no PhD’s.
How Google Works
Brin and Page realized that there was something important on the web that let them rank pages – hyperlinks. The links were simple ways for users to get from one page to another, but they realized the links could also be considered votes. The more incoming links a page had, the higher it would score in their algorithms, PageRank. They also factored in the importance of the linking pages, and the number of outgoing links each page had. This algorithm was finally able to rank pages in a meaningful way, and so Google became quite popular. Google also took into account the frequency and placement of different words, and has added many other factors into its analysis, but PageRank still plays an important role in their rankings.
Issues with Google’s Algorithm
An issue immediately arose with Google’s algorithm. A site could accrue many incoming links for two reasons:
1) It was a high quality sit, so many websites linked to it.
2) It spent a lot of effort and money into getting people to link to it so as to boost its PageRank.
The problem with PageRank is that it was too easy to be gamed. As Google became more popular, this became an ever-increasing problem. If only a few people used Google, it wouldn’t be worth the website owners’ efforts to try to game it. But once Google became so big, a site’s entire existence was defined by their rank on Google. Websites had to put a lot of effort into using the right keywords and getting lots of links, and many low-quality sites succeeded at this. Google tried to counter such moves with more advanced algorithms, but there was always a constant cat-and-mouse game between the websites and Google. In the past year, this issue has been raised in many blogs and newspapers. The New York Times published three articles on websites gaming Google searches (The Eyelasses Crook, JCpenney and Online Flowers Stores). Many topics on Google were returning low-quality sites derided as “content farms” like eHow.com and Associated Content as the top results. While these sites were sometimes useful, they often just stated the obvious without adding any real insight to the topic at hand. The content on the sites is created entirely so as to appear on Google results. Algorithms generate topics and titles based on Google searches and low-paid content writers quickly pump out articles to match the titles. Google’s algorithms were being beaten by content algorithms.
Like in quantum mechanics, Google was unable to remain an external obsevrer, their very act of looking changed the nature of the web.
Google responded to these criticisms with some algorithm updates, but with mixed results. It overall seems to have improved their quality, but low-quality sites still frequently come up high in the results. The problem has been especially problematic in certain high-expense advertising areas. Many sites that are mostly filled with affiliate links (where they get paid for each referall they send) have managed to reach the top of Google’s results with their “reviews” of the topics. For example, try searching for informative reviews of web hosts, online colleges or penny auctions.
Many sites will pretend to give honest reviews without any disclaimer, but really just contain empty positive reviews so as to get you to click on their affiliate links. On occasion, they will hide an honest disclaimer, such as this top-ranking site for the search “penny auction scam”:
“The website www.pennyauctionsreview.org receives compensation for the reviews posted on the site. All the reviews on this site are meant for entertainment purposes only and shouldn’t be taken as fact. Many of the reviews are in fact sent in via email from other people. PennyAuctionsReview.org does not own any of the trademarks used on this website. The typical person does not win anything by using penny auction sites.” Few visitors will end up reading their disclaimer, but I’m not sure this is the type of content someone investigating penny auctions is looking for. There are also searches where the top Google results have returned incoherent combinations of keywords without any meaning in English:
“With Purchase in every click advertising You could get Suited internet websites visitors into your Company net site The time and effort Your online Marketing is active.”
Other Issues with Google
Google still does not understand many types of queries, so even with a good algorithm for ranking, it won’t know what to search for. For example, if you type in a slightly complicated question, Google will often return irrelevant results. Often you will want to focus on a specific category or group of sites, but that is difficult to do on Google. There are times when the link-based algorithmic approach may not be able to tell which site will be most useful to an actual user.
Sometimes certain sites with overall good content always make it to the top of Google, even when some of their pages do not deserve to be there or are not the best for the user. For example, Wikipedia is on the top of Google for many types of searches. Yet often Wikipedia might not be written clearly enough for the average user who is not familiar with the topic.
Alternatives & the Future of Search
Google is constantly testing and improving their search, but there are also alternatives to Google. Bing is the only big competitor to Google, since they also power Yahoo’s search. Many of their results are similar to Google’s but they have more features for certain categories. Blekko is a recent entry to search which uses human editors along with their algorithms, and tries to only display results from high-quality websites. They also allow users to search specific categories of user-created topics. Google and Bing both recently added the ability to see friends recommendations in search results, and they might eventually allow user votes to affect global search results. There are also other tools a person can use to find information and websites besides a general search engine. Delicious.com and Xmarks lets you search through other people’s bookmarks. The old-fashioned directory, such as the wiki-style Open Directory can occasionally be useful to explore a certain category. If you have a question, Ask.com’s search engine is now focused on finding if anyone has asked it before. You can also try going directly to high-quality Q&A sites that cover your area, such as Quora or, for some topics, StackExchange. One of Google’s weakest areas also happens to be the source of much of their revenue. Google makes 99% of their revenue from advertising, but especially in certain categories. They get most of their billions from advertising in commercial categories, especially certain ones such as insurance, loans, and law. Yet these results are frequently low-quality. People are looking for help with major purchases, they need a site that will either solve their issue for them or point them to the right site consistently. If a company is able to help crack such an issue, they may be able to grab a piece of the billions in advertising money that is available. In the meantime, users can find other sites to help them with their purchases. There are many sites that help the user find the best deal within a specific area, such as the shopping comparison search Nextag, the compare-and-sort site FindtheBest, the review site ConsumerSearch or the bill-analysis site BillShrink. With sites such as these, the user can narrow in on the best deal at the best price.
In the 1990’s the Internet changed the world by providing a way for every type of information to be accessed from anywhere. Yet, this plethora of information would have been impossible to use without technology capable of sifting through it to find what the user needed. Many search engines tried different techniques to rank search results, but none of them did a satisfactory job.
In 1998, Google entered the scene. While Sergey Brin and Larry Page were graduate students at Stanford, they developed a new way to rank search results and created Google.com. However, Google soon became a distraction from their important work pursing a PhD, so they tried selling it for $1 million, but no one wanted to buy it. Sadly, they were forced to drop out of school to focus on their company. By 2007, their company was worth more than $200 billion, but they still had no PhD’s.
How Google Works
Brin and Page realized that there was something important on the web that let them rank pages – hyperlinks. The links were simple ways for users to get from one page to another, but they realized the links could also be considered votes. The more incoming links a page had, the higher it would score in their algorithms, PageRank. They also factored in the importance of the linking pages, and the number of outgoing links each page had. This algorithm was finally able to rank pages in a meaningful way, and so Google became quite popular. Google also took into account the frequency and placement of different words, and has added many other factors into its analysis, but PageRank still plays an important role in their rankings.
Issues with Google’s Algorithm
An issue immediately arose with Google’s algorithm. A site could accrue many incoming links for two reasons:
1) It was a high quality sit, so many websites linked to it.
2) It spent a lot of effort and money into getting people to link to it so as to boost its PageRank.
The problem with PageRank is that it was too easy to be gamed. As Google became more popular, this became an ever-increasing problem. If only a few people used Google, it wouldn’t be worth the website owners’ efforts to try to game it. But once Google became so big, a site’s entire existence was defined by their rank on Google. Websites had to put a lot of effort into using the right keywords and getting lots of links, and many low-quality sites succeeded at this. Google tried to counter such moves with more advanced algorithms, but there was always a constant cat-and-mouse game between the websites and Google. In the past year, this issue has been raised in many blogs and newspapers. The New York Times published three articles on websites gaming Google searches (The Eyelasses Crook, JCpenney and Online Flowers Stores). Many topics on Google were returning low-quality sites derided as “content farms” like eHow.com and Associated Content as the top results. While these sites were sometimes useful, they often just stated the obvious without adding any real insight to the topic at hand. The content on the sites is created entirely so as to appear on Google results. Algorithms generate topics and titles based on Google searches and low-paid content writers quickly pump out articles to match the titles. Google’s algorithms were being beaten by content algorithms.
Like in quantum mechanics, Google was unable to remain an external obsevrer, their very act of looking changed the nature of the web.
Google responded to these criticisms with some algorithm updates, but with mixed results. It overall seems to have improved their quality, but low-quality sites still frequently come up high in the results. The problem has been especially problematic in certain high-expense advertising areas. Many sites that are mostly filled with affiliate links (where they get paid for each referall they send) have managed to reach the top of Google’s results with their “reviews” of the topics. For example, try searching for informative reviews of web hosts, online colleges or penny auctions.
Many sites will pretend to give honest reviews without any disclaimer, but really just contain empty positive reviews so as to get you to click on their affiliate links. On occasion, they will hide an honest disclaimer, such as this top-ranking site for the search “penny auction scam”:
“The website www.pennyauctionsreview.org receives compensation for the reviews posted on the site. All the reviews on this site are meant for entertainment purposes only and shouldn’t be taken as fact. Many of the reviews are in fact sent in via email from other people. PennyAuctionsReview.org does not own any of the trademarks used on this website. The typical person does not win anything by using penny auction sites.” Few visitors will end up reading their disclaimer, but I’m not sure this is the type of content someone investigating penny auctions is looking for. There are also searches where the top Google results have returned incoherent combinations of keywords without any meaning in English:
“With Purchase in every click advertising You could get Suited internet websites visitors into your Company net site The time and effort Your online Marketing is active.”
Other Issues with Google
Google still does not understand many types of queries, so even with a good algorithm for ranking, it won’t know what to search for. For example, if you type in a slightly complicated question, Google will often return irrelevant results. Often you will want to focus on a specific category or group of sites, but that is difficult to do on Google. There are times when the link-based algorithmic approach may not be able to tell which site will be most useful to an actual user.
Sometimes certain sites with overall good content always make it to the top of Google, even when some of their pages do not deserve to be there or are not the best for the user. For example, Wikipedia is on the top of Google for many types of searches. Yet often Wikipedia might not be written clearly enough for the average user who is not familiar with the topic.
Alternatives & the Future of Search
Google is constantly testing and improving their search, but there are also alternatives to Google. Bing is the only big competitor to Google, since they also power Yahoo’s search. Many of their results are similar to Google’s but they have more features for certain categories. Blekko is a recent entry to search which uses human editors along with their algorithms, and tries to only display results from high-quality websites. They also allow users to search specific categories of user-created topics. Google and Bing both recently added the ability to see friends recommendations in search results, and they might eventually allow user votes to affect global search results. There are also other tools a person can use to find information and websites besides a general search engine. Delicious.com and Xmarks lets you search through other people’s bookmarks. The old-fashioned directory, such as the wiki-style Open Directory can occasionally be useful to explore a certain category. If you have a question, Ask.com’s search engine is now focused on finding if anyone has asked it before. You can also try going directly to high-quality Q&A sites that cover your area, such as Quora or, for some topics, StackExchange. One of Google’s weakest areas also happens to be the source of much of their revenue. Google makes 99% of their revenue from advertising, but especially in certain categories. They get most of their billions from advertising in commercial categories, especially certain ones such as insurance, loans, and law. Yet these results are frequently low-quality. People are looking for help with major purchases, they need a site that will either solve their issue for them or point them to the right site consistently. If a company is able to help crack such an issue, they may be able to grab a piece of the billions in advertising money that is available. In the meantime, users can find other sites to help them with their purchases. There are many sites that help the user find the best deal within a specific area, such as the shopping comparison search Nextag, the compare-and-sort site FindtheBest, the review site ConsumerSearch or the bill-analysis site BillShrink. With sites such as these, the user can narrow in on the best deal at the best price.
Tuesday, September 13, 2011
Computers Begin Taking Over
Soon computers will be diagnosing patients' problems and answering their legal questions...
Watson, the "Jeopardy!"-playing computer system, is getting a job... WellPoint said it plans to use Watson's data-crunching to help suggest treatment options and diagnoses to doctors. It is part of a far broader push in the health industry to incorporate computerized guidance into care, as doctors and hospitals adopt electronic medical records and other digital tools that can record, track and check their work.
http://online.wsj.com/article/SB10001424053111903532804576564600781798420.html
Monday, September 12, 2011
Posterous Platform Adjusts Focus to Photo Sharing
I've been blogging on the Posteorus platform for about a month and its worked OK. It is able to quickly turn emails into blog posts, so it does the basic job effectively. However, it is missing many features, such as the ability to add useful widgets to the side of the blog. In addition, Gmail does not make the best text editor, since its also a bit oversimplified. For example, its missing basic formatting options that exist in other programs.
Today, Posteorus just launched a new focus on being a kind of social network and photo-sharing site. The blogging platform will still exist, but they will no longer be focused on blogs. So I guess I won't be able to expect any blogging platform imptovement. I wish them luck, but I will probably move this blog to Wordpress soon.
News sites:
Saturday, September 10, 2011
Teaching through Games
In this Quora answer, I suggested a few ways that games can be used to teach advanced subjects. For example:
[To learn] programming, one can be required to create programs to complete missions.
I recently came across a mention in the WSJ of a game that will do exactly that. From the website of CodeHero:
It is a co-op first-person science shooter where you use the code gun to shoot javascript. Your code gun can copy code like new items and fire it like ammunition to do new things.
You can edit new code to do anything you can imagine. You'll learn how to blast the enemy, manipulate the world, and build structures creatively to create the games of your dreams and recruit an army of coders to save the world from rogue AI.
In the meantime, there's always programs like Greenfoot, Scratch and Microsoft SmallBasic to teach kids programming.
Monday, September 5, 2011
Guide to Building Websites - No programming required - Part 1
Introduction
Many people want to build websites but think they will have to learn programming to do so. This series of posts will show you how to build a website without requiring any coding. Many basic sites can be set up without writing any new code at all. However, programming will be necessary if you want to customize things further, or to create a new type of site. This series will therefore focus on solutions that will work well with actual custom coding later. These posts will also be useful for someone with some programming experience who wants to learn the non-programming aspects of building websites.
What is a website?
A website means you have files hosted on a computer that is always connected to the internet, which is known as a server. The server sends webpages over the internet so that anyone connected can view them. So when a user goes to Amazon.com, a server from Amazon send his computer Amazon's homepage. Technically, any computer can be turned into a server, but practically, you will want to use a web host to host your files on their servers.
Many people want to build websites but think they will have to learn programming to do so. This series of posts will show you how to build a website without requiring any coding. Many basic sites can be set up without writing any new code at all. However, programming will be necessary if you want to customize things further, or to create a new type of site. This series will therefore focus on solutions that will work well with actual custom coding later. These posts will also be useful for someone with some programming experience who wants to learn the non-programming aspects of building websites.
What is a website?
A website means you have files hosted on a computer that is always connected to the internet, which is known as a server. The server sends webpages over the internet so that anyone connected can view them. So when a user goes to Amazon.com, a server from Amazon send his computer Amazon's homepage. Technically, any computer can be turned into a server, but practically, you will want to use a web host to host your files on their servers.
So a website is just a collection of files?
The easiest way to create simple website is to use a DIY-service such as Google Sites or Weebly.com. They offer easy-to-use interfaces for creating simple sites. However, if you want to be able to customize more aspects of your site and have more control over the actual HTML, you'll need a web host that lets you upload your own web files. This guide will assume you will pick a host with such flexibility. There are many free options you can choose from, such as Webs.com, Microsoft Office Live or Static Cloud.
This depends on what type of site you have - static or dynamic. A static site doesn't do anything, it can just display text, images, links and videos. The files are stored on a server and sent over to the user when requested. This may be enough for your site if all you need is a simple online presence for yourself or your business. The pages just consist of simple pages made with HTML and CSS. You can't run any code from the server, but you can use javascript to run simple code from the users' computers. However, if you need a site that can 'do stuff' such as using databases and running code, you'll need a dynamic site, which will be discussed in a future post.
The easiest way to create simple website is to use a DIY-service such as Google Sites or Weebly.com. They offer easy-to-use interfaces for creating simple sites. However, if you want to be able to customize more aspects of your site and have more control over the actual HTML, you'll need a web host that lets you upload your own web files. This guide will assume you will pick a host with such flexibility. There are many free options you can choose from, such as Webs.com, Microsoft Office Live or Static Cloud.
What is a Domain Name Registrar?
A domain name registrar is where you buy your domain name from, such as Example.com. A website actually requires two things - a web server that holds the websites, and a domain server that sends people to the web server when they type in the website name. For example, when you type in Amazon.com into your browser, the name is sent to a Domain Name Server which looks up Amazon.com and then sends your computer the information to connect to Amazon's web servers. The web hosting and domain hosting are often packages together, but they're two separate aspects. So you can buy your web hosting from one place, and then buy five domains from another place.
Where should I buy a Domain Name?
If you're building a simple site, Doteasy.com offers a basic web hosting package with your Domain purchase. If you picked a web host separately, they normally also offer the option to buy a domain name. Though if you want a special deal, GoDaddy.com and 1and1.com often offer domain names for $1 for the first year (You can check deals here). If you're going to get a dynamic host, most of them will throw in a free domain name. Otherwise most Domain Names cost about $10 a year. After you have a domain and a web host, you'll be able to create static websites.
How will I do that?
Stay tuned for the next post...
The Modern Doctor's Office
Last week, I posted that the health sector hasn't modernized and isn't transparent. When ever you visit a new doctor, you have to fill out tons of the same paper forms, and its difficult for one doctor to get information from another. In addition, the doctor has to pay various secretaries and administrators to handle all the paperwork. Today, one startup has raised money to help improve things:
One Medical operates 9 doctor’s offices in San Francisco and New York, and will open 5 more this year, expanding to Silicon Valley and Washington, D.C. Patients can schedule appointments online, request prescriptions, get lab results digitally, and see their personal health summary online. Doctors can access medical records electronically (One Medical designed its own electronic medical record with doctors and patients in mind, not administrators). One benefit of having digital medical records is that patients can visit any office since every doctor has access to their records.
New patients can join online, and pay online. It even has its own iPhone app for scheduling appointments. Simple questions which can be addressed via email or the iPhone app are done digitally instead of requiring an in-person visit. And when patients do go in, the offices are bright, airy and modern.
http://techcrunch.com/2011/09/05/one-medical-raises-20-million/
Sunday, September 4, 2011
How to claim your Blog on Technorati 3ZP97E2ENKSW
3ZP97E2ENKSW Technorati might serve some use, so it makes sense to claim one's blog on it. You create an accoutn, submit your blog to them, and then post the code they give you. In this case: 3ZP97E2ENKSW
Guide to Selling Your Textbooks
If you're a student, you both need to buy textbooks and figure out what to do with your old ones. In most cases, you probably won't look at the textbooks very much in the future, so you should sell them quickly before they depreciate in value. The textbook publishers try to come out with new ones every few years so as to keep sales high. Once the new edition comes out, the prices of the old ones drop dramatically, so don't delay.
Where to sell Textbooks
Another Student - Again, this is the simplest option if you can do it. You avoid paying fees to any site, and you don't need to bother shipping the book anywhere. You can post an ad on college classifieds site, or put up a sign in the classroom that will have the same course that you took. Then you meet the buyer, exchange the book for cash, and you're done, without any fees involved.
Amazon or Half.com - If you can't find another student, this is second-most profitable option. You enter the ISBN and condition of your book and it gets listed on the site. Amazon is the most popular site for buying books, so its probably your best bet. However, Half.com has slightly smaller fees, so it might be worthwhile to sell popular books on it. These are the exact fees:
Both sites charge 15% of the purchase, but Amazon also charges an additional flat-rate of $2.30 for every book sold through their site. Meanwhile Half secretly takes part of the shipping commission, so you only get $2.64 or $3.07 as a shipping credit, depending on whether its soft or hard-cover. However, Half does lower the fees for books over $50.
After the book sells, you need to ship it to the buyer within a couple days. Normally, the cheapest way to to this is with media mail. Both Half and Amazon let you buy shipping labels through their site, which is the most convenient option. A single book normally costs $2.41 - 3.64 to ship, depending on the weight. You then print out a shipping label, stick it on a padded envelope, and mail the book away.
This table summarizes the costs of each site:
| Amazon.com | Half.com |
Commission charged | Price * 85% – 2.34 fee + 3.99 shipping credit | Price * 85% + 3.64 (hardcover) Price * 85% + 2.07 (softcover) |
Money received on a 2-pound hardcover $30 Textbook after shipping costs. |
Buyback Sites - Another option is to sell your book to a buyback site. This normally isn't worth it though, since they pay much less than if you just sell it yourself, and you still need to ship it somewhere. The exception is in the rare case where you get a decent amount through these sites. The book must be in very good condition though. As with buying books, you can compare buyback prices with a price comparison search, such as on DirectTextbooks, or NYtextbooks. Amazon Buyback occasionally offers a decent amount of Amazon credit for a book, so that can be a good option if you'll anyways be buying from Amazon in the future.
Campus Buyback - This is the most convenient option, if its offered on your campus. Be careful not to get ripped off though. When I ran a buyback, I payed 60-70% of the book's used selling price on Amazon. Amazon than took 15% and other fees. Then there were other costs, such as for envelopes and barcode scanners, and some books didn't sell, so I had a small but reasonable profit in the end. Just check the price on Amazon before selling your book and make sure you're getting over half of it. Otherwise, its worth it to just sell the book yourself online.
Friday, September 2, 2011
Google Kills 10 Products
Over the next few months we’ll be shutting down a number of products and merging others into existing products as features. The list is below. This will make things much simpler for our users, improving the overall Google experience...
Here’s a quick overview of where a number of products and features are headed:
- Aardvark: Aardvark was a start-up we acquired in 2010. An experiment in a new kind of social search, it helped people answer each other’s questions. While Aardvark will be closing, we’ll continue to work on tools that enable people to connect and discover richer knowledge about the world.
- Desktop: In the last few years, there’s been a huge shift from local to cloud-based storage and computing, as well as the integration of search and gadget functionality into most modern operating systems. People now have instant access to their data, whether online or offline. As this was the goal of Google Desktop, the product will be discontinued on September 14, including all the associated APIs, services, plugins, gadgets and support.
- Fast Flip: Fast Flip was started to help pioneer news content browsing and reading experiences for the web and mobile devices. For the past two years, in collaboration with publishers, the Fast Flip experiment has fueled a new approach to faster, richer content display on the web. This approach will live on in our other display and delivery tools.
- Google Pack, Notebook, Sidewiki [and 4 more...]
http://googleblog.blogspot.com/2011/09/fall-spring-clean.html
Pretty Crazy. I occasionally used Aardvark if I needed a quick answer to a question. What other service emails people questions and send back answers so quickly? I would've thought Google would have tried to incorporate it more into Google+ to compete with Facebook Answers. FastFlip was also an interesting way to browse news, though I don't know if it was actually that useful. And Google Desktop was a good way to search one's computer, though Windows has improved greatly on that front. I don't think they'll actually be able to get everyone to move to the cloud that quickly, but I guess they want to stay focused.
Thursday, September 1, 2011
Why Governments Don’t Get Startups « Steve Blank
To date, Israel is only country that has engineered a successful entrepreneurship cluster from the ground up. It’s Yozma program kick-started a private venture capital industry with government funds, (emulating the U.S. lesson of using SBIC funds.), but then the government got out of the way.
In addition, the Israeli government originally funded 23 early stage incubators but turned them over to the VC’s to own and manage. They’re run by business professionals (not real-estate managers looking to rent out excess office space) and entry is not for life-style entrepreneurs, but is a bootcamp for VC funding.
via steveblank.com
Government and private enterprise don't generally go well together, but Israel managed to get venture capital flowing by gathering the money and then getting out of the way.
Subscribe to:
Posts (Atom)