MetaScroll Search Blog

Google will follow the activities of users across e-mail, search, YouTube and other services

Wed, 25 Jan 2012 00:31:14 +0100

Google will follow the activities of users across e-mail, search, YouTube and other services

Many Google employees seem to be unhappy about Google's new SERPs.

Wed, 25 Jan 2012 00:18:35 +0100

Many Google employees seem to be unhappy about Google's new SERPs.

Google puts it's users in a search bubble, making climate war worse.

Wed, 02 Nov 2011 12:17:19 +0100

Google puts it's users in a search bubble, making climate war worse.

Links to Wikipedia are now protocol relative

Tue, 04 Oct 2011 00:58:02 +0200

MetaScroll now supports Wikipedias protocol relative URLs. So when you are using an SSL encrypted connection to MetaScroll an go to a Wikipedia article from there, it will be SSL encrypted as well.

No chance for any possible listeners along your internet connection (educational facilities, employers, café owners) to see what you were searching or what Wikipedia articles you were reading.

Google Admits Handing over European User Data to US Intelligence Agencies

Thu, 11 Aug 2011 15:24:16 +0200

Google Admits Handing over European User Data to US Intelligence Agencies

Yahoo! now stores their log files for 18 instead of 3 months

Mon, 18 Apr 2011 20:38:34 +0200

Yahoo! now stores their log files for 18 instead of 3 months

btw: The quick preview feature has a keyboard shortcut

Fri, 25 Mar 2011 22:36:17 +0100

I totally forgot to mention: the shortcut for the web preview is Ctrl+Enter. Have fun!

Anonymous peak into search result pages

Tue, 22 Feb 2011 23:51:10 +0100

Starting now, you find an new icon in the upper right corner of every search result on MetaScroll:

By clicking this icon, you can get a quick preview of the page behind the link, without ever visiting it:

This helps you find the right site for your search, without revealing your IP to different websites and multiple ad/tracking providers along the way. Try it out!

Technical details

This feature is implemented using the freshly available PhantomJS software. PhantomJS is basically a headless Webkit browser, which can be controlled using JavaScript. Wrapped inside a special little NodeJS webserver (for some caching and stuff), PhantomJS is used to generate the screenshot for a given URL, which then is delivered to the user by NodeJS. This sure will be the starting point for many other great features. Stay tuned!

Wallstreet Journal: Rivals Say Google Plays Favorites

Tue, 14 Dec 2010 10:11:50 +0100

Wallstreet Journal: Rivals Say Google Plays Favorites :

Good for those, who don’t rely on Google alone.

Viglink integration

Sat, 13 Nov 2010 00:36:06 +0100

I just integrated the Viglink-service as an intend to cover the expenses necessary to support MetaScroll.

What Viglink does is, it tries to find an affiliate program for a given URL. All clicks on a search result on MetaScroll are now redirected via Viglink as a default. You an prevent this by unchecking the “Allow affiliate programs”-checkbx in the settings. Viglink takes the given URL, checks it against an internal database of given affiliate programs and performs a redirect via this program to the according URL.

MetaScroll does not give any personal data to Viglink (how could we, MetaScroll doesn’t store any personal data at all), but Viglink will save some anonymous identifier on your browser, to be able to determine a generated value, on which MetaScroll can participate.

I think the unintrusive, anonymous character of this monetisation is a good match for the MetaScroll service. And as already stated, you can always opt out of it.

Google reveals user search histories 9k times every year & Eric Schmidt says you should keep your privacy private!

Wed, 27 Oct 2010 14:10:58 +0200

Google reveals user search histories 9k times every year & Eric Schmidt says you should keep your privacy private!

Switch to Google Closure

Mon, 25 Oct 2010 23:37:14 +0200

MetaScroll went through a major rewrite of the search client during the last weeks. I switched from using jQuery as the core JavaScript library to Google’s Closure Library.

The reason I chose Closure is the hope for a better overall performance due to less encapsulation, as well as a smaller footprint of the JavaScript Code.

The Google Closure Library allows one to pick the required modules according to one’s needs, quite similar to the import or require routines known from other languages like Java. The resulting client code is a lot larger than the “old” jQuery using version, due to the more explicit coding style of Closure. But this effect changes drastically after the compilation of the JavaScript source code using the Google Closure Compiler with its ADVANCED_OPTIMIZATIONS-setting: The compilation process reduces the size of the code base with all its dependencies from ~815kB to ~55kB! If you want to learn more about how the compiler achieves this reduction you can find a brief overview in the documentation. After all, the rewrite resulted in a size decrease of around 40kB, from ~95kB incl. jQuery to now ~55kB with Closure.

During the rewrite I also reworked some of the older parts of the code, to aid and increase the speed of future developments, which you will get to see quite soon! If you have an idea or feature wish as well, just let me know!

MetaScroll is "Mashup of the Day" on ProgrammableWeb

Thu, 02 Sep 2010 10:29:28 +0200

ProgrammableWeb chose MetaScroll as their “Mashup of the Day” for August 30th 2010! This is the first coverage of MetaScroll on international websites.

Thank you very much, guys! Great to see that others love MetaScroll as much as I do!

Keyboard support!

Thu, 05 Aug 2010 23:28:00 +0200

Your mouse or trackpads are great ways to skim through your search results on MetaScroll! Anyhow, some folks prefer to not travel with their hands between keyboard and mouse/trackpad when searching. They’d ike to leave them on the keyboard.

For those keyboard afficiandos MetaScroll just got a lot better! You now aren’t dependend on your mouse anymore! Instead, you can move between search results using your cursor keys. To open a search result, just use “Enter” or the “Space”-bar.

Folling you’ll find a list of shortcuts:

With the searchbox unfocussed:

Just start typing for a new search, the searchbox gets focussed automatically
You can hit “Esc” to focus the searchbox, without overwriting the previous search term

With the searchbox focussed:

“Enter” performs a search, selecting the first search result
“Up”/”Down” switches to the next/previous search suggestion
“Tab” unfocusses the searchbox and selects the first search result, without doing a search

Moving in search results:

“Up” or “k” select the previous search result (YES! I like vi, too!)
“Down” or “j” select the next one
“Enter” or “Space” open the currently selected result

You may have to allow popups for the domain www.metascroll.com to be able to open a search result using the keyboard. These get opened in a new window or tab, depending on your browser setting.

I really hope you like these new features as much as we do! If you have any suggestions for other shortcuts, please, don’t hesitate to let me know!

New: Now with MIME-type hinting

Tue, 16 Mar 2010 01:11:40 +0100

I am proud to annonce that by now a new feature just got released! MetaScroll now hints various non-HTML MIME-types to you by displaying an according icon (taken from the great KDE Oxygen set).

You can try it yourself by doing a search for e.g. some PDFs. Hinted MIME-types include, but are not limited to:

PDF
MP3, WAV, AAC
WMV, AVI, MOV
DOC, XLS, CSV, RTF
TTF, Torrent, Diff

If you have a MIME-type you’d like to see included, please drop me a line!

Webserver configuration

Mon, 08 Mar 2010 22:49:06 +0100

At first, I used the classic allrounder, the Apache2-webserver. But since its memory footprint isn’t the smallest, especially with all the modules I use (like mod_deflate, mod_jk, mod_proxy, mod_php, mod_rewrite, mod_ssl, etc), a service like MetaScroll generates quite a load on the machine, even with very few users. This is because of many requests being proxied over the service, so that only one search generates at least three requests on the server, not counting the suggestions.

The solution I chose was the use of the lightweight nginx-webserver. It isn’t as full-featured and modular as the Apache, but it comes with everything I need and a lot more performance while using only a percentage of Apache’s memory.

You can find the relevant settings I use below.

Nginx

Gzip compression
gzip on; gzip_http_version 1.0; gzip_vary on; gzip_comp_level 6; gzip_proxied any; gzip_types text/plain text/css application/json application/x-javascript text/xml application/xml application/xml+rss text/javascript; gzip_buffers 16 8k;
Turn off server tokens
server_name_in_redirect off; server_tokens off;
Various TCP-settings
sendfile on;
tcp_nopush on;
tcp_nodelay off;
keepalive_timeout 10;

The technology behind MetaScroll

Sat, 20 Feb 2010 01:32:00 +0100

In the cause of this blog, I will try to give you an impression of the technology behind MetaScroll and the ropes that I’ve jumped through, to give you a top performance for your we search. Hopefully you can learn a thing or two, to help you with YOUR projects.

The topics that I’ll try to cover are …

Caching
Document size reduction
Webserver configuration
Backend performance
Frontend performance

So stay tuned for the things coming up!

What is MetaScroll?

Sat, 20 Feb 2010 01:22:36 +0100

MetaScroll is the attempt to build a modern version of the classic meta-search engines. These engines, though quite popular in the late 90s, early 2000s, lost their market share to Google, like all the other of the popular search engines at that time.

Although Google delivers great results and a good searching experiences, there still are issues with their web search:

Ease of use. At the end of a result page, one has to manually find and click the “next”-button to get to the next page of web results. This process interrupts the visual scanning of the results.
Search performance. While Google’s search result pages are undoubtingly fast, having to find the “next”-button and click it to get to the next page is a repetitive and slow task.
Direct and relevant answers. Google does an incredible job finding the best websites regarding your search. But you still have to open the page with the information you’re looking for. And the info on this page could be days, weeks, or even years old when you are searching for something thats happening right now!
Privacy. This is the most important issue. Google does not only store every search you make and when you make it, it also stores which results you click on and if or how much later you come bck to their site. The collected information, together with the data from other Google services like Mail, Calendar, Reader, etc lead to a very distinct profile of you and your interests.

Of course you’re free to use other search engines like Microsoft’s Bing or Yahoo!, but the issues named above are valid those as well, to a greater or smaller extend.

So, how does MetaScroll fit into the picture? What can it offer regarding the issues named above?

Ease of use. There is now “Next page”-button on MetaScroll you’d have to find and point your mouse at. The result just keep coming while you scroll down the page.
Search performance. The process of scrolling down the page doesn’t get interrupted with new pages. MetaScroll delivers the results in one large list that grows while you scroll.
Direct and relevant answers. Classic search engine results are integrated with relevant Wikipedia articles and other people’s messages from Twitter. This way you get a 360° view on the topic you’re searching for.
Privacy. MetaScroll does all it can to ensure that no search engine has the possibility to identify you while you’re searching. And of course no personal information is stored by MetaScroll itself

.So you see, MetaScroll gives a great deal about improving your searching experience. And if there is anything you see that could fit MetaScroll and it’s philosophy, or that you would need to definitly do the switch from yourcurrent favorit search engine, you’re more than welcome to get in touch!

Let me know your thoughts!

Hendrik, Developer of MetaScroll.com