|Google Launches Search Service For Computer Code|
|Google Inc. is introducing a new search service that only a geek could love. |
The Web search leader said late on Wednesday it is introducing Google Code Search, a site that simplifies how software developers search for programming code to improve existing software or create new programs.
Google product manager Tom Stocky said the Mountain View, California-based company is set to help programmers sift through billions of lines of computer source code using its familiar search box to uncover snippets of reusable software.
Google Launches Search Service For Computer Code [today.reuters.co.uk]
It's in Google Labs at the moment.
The buzz is that "Just by searching for "keygen", "serial", "name", and some well known cracker groups you can come by the keygen sources (how serial numbers are generated) to some very popular programs. I found WinZip, Photoshop, mIRC, and a few others."
Other troubles in sight?
|you can come by the keygen sources |
Whoa, I never thought about that...very interesting. I'm more excited about finding clever implimentations of regular expressions. This is a great day for programmers everywhere!
Regarding the "keygen" insight, I've begun to wonder how well the creators of all the new mashups, webservices, and assorted webapps think through the possible unintended or malevolent uses and bad outcomes of their actions.
In the case of "things released on the WWW' it ought to be known by now that "If you built it and release it they (bad guys) will come and use your webapp for their own benefit (detriment to others)".
Will people, legilatures, courts, victims, the world eventually tire of the statement "Well, we didn't build our software/webapp/online service for that purpose. We can't control how people use it."
Guns and webapps: Are they the same or are they different? Are WebApps WOMD, unlike guns?
[edited by: Webwork at 3:18 pm (utc) on Oct. 5, 2006]
Yep, I can find some of my own code in there <specfiics removed> which was a bugger to write! This means that hopefully a few other people can crib my code and be saved the pain!
<Sorry, no specifics.
See Forum Charter [webmasterworld.com]>
[edited by: tedster at 3:10 pm (utc) on Oct. 5, 2006]
Regarding the "keygen" thing - who cares?
This can already be searched for on Google and any other search engine just as easily if you know what your doing.
Why not start a seperate thread about how "keygen code" can be searched for on ALL search engines?
Back to the topic on hand...
I think that this new code search facility is great and I'm hopefull that it will save me lots of time when I'm coding in php/asp/c. Thanks Google :)
From a c¦Net article:
|The search engine will let people do both keyword search and "regular expressions," which allow people to search a specified pattern, he said. |
For me, THIS is the big news here.
I'd like to see them apply regular expressions (or similar concepts) elsewhere.
For example, in Adwords. It's rediculous that advertisers have to specify keywords exhaustively. I want to be able to say "(red¦green¦blue) (widget¦thingamabob)s?".
This is a great service but Google are missing a trick.
Searching for ::= in Google code finds plenty of uses, but not one explaination of the meaning of ::=. Try the same search in the main Google index and get not a single result.
How is one supposed to find out what ::= means, if one does not already know?
That they are also indexing .tex files (although it is not on the list of languages under Advanced Search). This might really save time entering complex formulæ.
There is a significant risk here because if a given function or method is revealed as exploitable for one given software package it will now be very easy to search through countless application source-codes for that same function and target them as well.
New Ideas for Use
This facility would be a great starting point for countless academic papers by computer scientists upon the actual usage of programming languages.
This could also be used to crack down upon the reuse of licensed code without permission.
A 'Translate this page' document with options for moving to a number of other languages as well as pseudocode.
A CVS type log / diff tool / some kind of comparative feature to show changes in code since previous versions.
Warnings where unsafe or dangerous code is shown. I've seen snippets there which will format your harddrive as well as the source code for email worms. It would be nice to be warned before I misunderstand what the code does and execute it.
I think codase has a better service. Some very good search options for Source code.
IMO, this service should require a positive permission from the site owner. There is just too much potential for abuse.
Google also needs to take some proactive steps to prevent searches for the purpose of finding passwords, credit card numbers, admin pages, etc.
There's just WAY too much stuff with potential for harm that you can find using Google. While it would be impossible for Google to exhaustively catch all cases, many of them would be EASY to implement.
There is no reason to index: control panel login pages, router/other appliance login pages, lists of credit card numbers (easily identified - hey, Google lets you search specifically for: UPC codes, shipment tracking numbers, etc. routing you to the right place. Surely, they can identify a credit-card number) etc. etc.
BTW, I see that Google tries to determine the license terms. But MOST entries indicate "Unknown license". The index purports to contain "open source" code. If the license is unknown, how to they know it is open source?
I think Google has moved dangerously close to leaving the cover off the pool, without a fence or an alarm. Maybe that's not enough for liability. Adding a sign "free ice cream here" might just be enough, though.
G managed to correctly identify the BSD-style licence that my code is under, and even dug it out of a BZIP2-ed tar archive.
Someday they will release search thru html code option
One interesting thing is the way in which it happily indexes documents within archives. I wonder when we'll see that in the main search index?
e.g. a zipped file of PDF papers might have index entries for each paper, the link going to a cached copy with a second link to download the archive.