12 August 2006
The Search for Secret Google Services
http://blogoscoped.com/archive/2006-08-11-n67.html
Since stumbling across Google Base almost a month before its release and more recently uncovering some more secret Google services, several people have asked how I go about finding these new services. By way of a response, here’s a quick guide to just some of the many different ways you could use to try and find new services that Google are working on.
1. Finding a new service
The details of new services can sometimes be leaked accidentally. Other times, they’re kept secret right up until their release. Here are a few things to look out for if you want to discover new services before anyone outside the Googleplex has even heard about them:
- New domain names
While Google tends to use their google.com domain to host new services, they often register new domain names well in advance of new services being released in order to stop domain cybersquatters snapping them up first. Gary Price has made a couple of posts about this recently. Whilst reverse DNS lookups can help, don’t get fooled into thinking that Google owns all the domain names that are resolving to its IP address ranges or servers. New domains can sometimes give useful hints about new services but can’t be relied upon. (Google has owned GBrowser.com for years, yet we still haven’t seen an official release.) - New subdomains
When I found the base.google.com subdomain, I simply used a script and a large list of around 65,000 English words to send requests in the format *.google.com to see which ones resolved. It’s probably also worth checking subdomains in the format of *.l.google.com as sometimes these are active well before the normal subdomain, as Garett Rogers pointed out when he found the writely.l.google.com subdomain was resolving back in June 2006 (whereas the writely.google.com subdomain still doesn’t resolve at the time of writing this). - New subfolders
Google has also been known to just use subfolders instead of subdomains when launching new services, including Google Notebook. Using a similar word list technique as I did for sniffing out subdomains, you could query all Google’s domains and subdomains for new subfolders. Be warned though – if you send too many requests to Google’s servers in a short period of time, you may get temporarily blocked! - New Google Account service names
Back in May, I tried a new approach to finding new services. I’d noticed that services requiring a Google Account used a series of similar URLs for creating new accounts and signing in. By appending various words to the end of these pages – i.e. https://www.google.com/accounts/Login?service= – I was able to find several new services such as Google Weaver (known back then as M Scrapbook), Google RS2, SSD, Mobile Download Console and LH2 (now known to be Picasa Web Albums). More recently, I tried a similar approach with the sandbox.google.com version of Google Accounts and found even more services, which Google swiftly removed. (I try to maintain a list of these Google Account service names on my website.) - Clues in JavaScript / HTML code
Another technique often used by Garett Rogers involves sifting through the JavaScript source code used by Google services such as Gmail and Google Calendar. This can reveal new domains, subdomains, links, labels and other hints at new services or features for existing services. - Changes to Google’s robots.txt files
Google try their best to not index their own SERPs, so when they create a new service that could clutter up their index, they’ll usually add a new entry to their robots.txt file. (One of the most recent additions to the file was www.google.com/call, as Garett Rogers discovered.) By scheduling a simple script, you could easily monitor this file and be alerted of any changes. Don’t forget that each Google domain and subdomain may have its own robots.txt file, so to do this thoroughly you would need to monitor all those for changes too! - Changes to Blogger profile pages
Since Google don’t use standard naming conventions for their official blogs, it’s hard to detect when a new official blog has been created. One of the best (if not only) ways to check if a blog is official is to see whether it appears on any Blogger profiles belonging to Googlers (such as Eric Case, for example). Quite often, the titles of new official Google blogs appear on these profiles pages before they (or their related services) are officially released. Monitoring these pages for changes could give clues about the names of any new services.
2. Waiting for the new service to be released
Once a new service is discovered, everyone always wants to be the first to confirm a sighting and post exclusive screenshots. Here’s what you can do if you’re expecting a service to be released:
- New logos and images
Sometimes it’s possible to discover new services by making educated guesses about the image and logo URLs. Logos often follow similar naming conventions, which makes it easier to guess the location of a new service logo once you already know the name. (This technique can also be used to find new Google holiday logos.) - New support pages
Sometimes Google puts the support pages online before releasing a service. Since the Google Help URLs often follow similar conventions, it’s relatively easy to check a few URL combinations to see if the new service’s help pages are online. - Updates to other pages
Sometimes, Google starts rolling out new services by quietly adding links to various pages, in particular Google Labs, the More Google products page and the Google Site Map page. By monitoring these pages for updates, you could be one of the first to know when a new service has been released, meaning you could be one of the first to sign up (before the number of signups get restricted) and get those all important screenshots if the service is taken offline again!
3. What to do once you find something new
When Corsin Camichel snooped around the Google servers, he accidentally hit on Google Platypus aka Gdrive (and took a screenshot, of course).
Even if you’ve not followed the previous steps, Google sometimes adds references and links to pages either accidentally or without announcing them. This is what you should do if you see something new:
- Take plenty of screenshots
Google has been known to make new services and features live for short periods of time before removing them again. If you find something new, make sure you take screenshots of all the pages you can find in case the pages disappear again. - Make copies of the source code
Screenshots are great but they only show you what’s on the surface. By saving the HTML and JavaScript source code, you will be able to search for any references to other new services even if Google take the new service offline. (Analyzing links and code for new services shortly after release can also sometimes reveal “secret” URLs that were meant only for testing and these can be useful in discovering new services in the future.) - Let others know about what you’ve found
No matter how thorough you think you may have been, someone else will probably be able to find something you didn’t. By sharing your findings, you can help others find out even more about the new services. If you have a blog, make a post about your findings including any screenshots you managed to grab. (And don’t forget to email Philipp Lenssen or create a thread in the Google Blogoscoped Forum!)
And here endeth the lesson. Others will probably have their own ways of doing things and new methods will emerge all the time.
Of course, you could just try bribing a Googler or one of their ‘Trusted Tester’ friends... [added: Just kidding!]
Labels: blogoscoped, google
