My Startup project: ErrorKey.com, a tool for hackers. | Hacker News Reader

My Startup project: ErrorKey.com, a tool for hackers.

errorkey.com

60 pointsokeumeni51 comments

I like the idea! I like it so much that our startup last year targetted the same general audience with a custom error search engine:

http://bug.gd

http://bug.gd/info/faq

We've only grabbed error codes where we can find/recommend a solution, but when it comes from crawlers we always link to the source as well to avoid copyright issues. How are you working around that in your case? Are they coming from local sources rather than websites? (Admittedly, I doubt these companies are too worried about their error documentation being quoted in full.)

Best of luck to you! I like to see more attention to this area, no doubt.

Despite the benign assurances of the site, having to type my email address in bug.gd is annoying. I disagree with "forcing community participation" like this... how do I even know there would even be enough content on your site to make it worth my time?

As much as I love hearing feedback (thanks!), I don't really want to get a side thread going about our site in this other team's thread. Not really fair to them.

As briefly as I can: The Firefox extension allows anonymous search/contribution. Stay tuned for changes more to your liking in our next site release.

You could suggest that they change their approach in getting the email address but the concept of reminding you to reply with your fix to that error is good.

I completely agree, its a good feature if its optional.

Is it going to be better than just putting the code in a search engine like google or something. Most of the codes I come across are unique enough that I get a direct match pretty quickly. Also what helps me is seeing related forum posts and other comments on resolving the issue, the definition in on itself is often useless.

Depending on the system your working on finding a fix or explanation to an error is not always easy. The next step is to allow users to post alternative fix to errors or link to forum where research can be extended.

I have nothing to do with the site or author, but I can suggest the following:

-Google defaults to under-emphasizing results related to errors specific in newer versions of software. So when a bug in a new release comes out, info is often no where to be seen--it is obscured by older items with higher rank.

-Google includes links to both questions, and answers. I only want answers.

-Google's results are across multiple pages and obscured in forum discussions. I want one page with all the answers and info.

-Google frequently returns results from experts exchange, which involves scrolling to the bottom. Additionally, I've talked to developers who don't even realize the unobscured answer is at the bottom.

-Some code queries do not translate well into regular searches because the characters have meaning. Off the top of my head example: select * from blah

-There is not always a nice centralized place to go to share it with the world know when you have a fix.

That said, I'm not sure this site implements solutions to all these problems, but a wiki-type site could. Isn't Jeff Atwood doing something with a similar idea?

I've been using Google for this for a very, very long time.

Can you describe to me why I would want to come to your site next time I had an error?

Error codes, without all the other crap (bullshit paid forums, AOL nation posts) is reason enough for me.

Absolutely. I can't tell you how friggin' sick I am of inadvertently clicking on an "Experts Exchange" link in a google search set. The microsecond it takes my eyes to register an EE site and click back is one too many. I'm pretty sure EE owes me about 10 years of my life up to this point.

Member-only EE content is visible in Google's cached version of their pages. Just scroll to the bottom part of the page to see it. E.g.:

http://209.85.141.104/search?q=cache:gRw9sjavhVAJ:www.expert...

It's available from the regular version of the page too--just scroll to the bottom like you do on the cache:

http://www.experts-exchange.com/OS/Microsoft_Operating_Syste...

Not for me. This 'trick' stopped working for me about a month ago, so now I have to use the cache links.

Interestingly it seems to be based on a referrer. It worked when I clicked the link from Google, not when I clicked it from here. Why are these clowns still in Google's index anyways, cloaking isn't allowed.

No, it doesn't work. Real answers appear in "collapsed" form, so they are not visible.

It's contingent on a Google referrer. They should be removed from the index.

Compare these:

  curl -e http://www.experts-exchange.com/OS/Microsoft_Operating_Systems/Windows/XP/Q_21692879.html?qid=21692879 > without_google.html

  curl -e http://www.google.com http://www.experts-exchange.com/OS/Microsoft_Operating_Systems/Windows/XP/Q_21692879.html?qid=21692879 > with_google.html

At the risk of getting downmodded, can I suggest that if you're getting helpful answers off EE, it would be nice to pay the subscription fee and help support the business?

It is EE's choice to play sleazy games where they represent one thing to search engines and deliver something else to the users who arrive at their site.

They can stop deceiving people about their public content and go to a standard gated subscription model any time they like. Until they do, the more people who work around their weasel tactics, the better.

It's a freemium model in disguise.

It's fair to be irritated by the search engine listings.

I'm going to disagree that it's equally fair to swipe content that you find to be helpful from a subscription-based site without paying for it.

I guess I see it like this: I'm irritated that there's a Starbucks on every corner of the SF financial district; but that doesn't mean I go in and grab a cup of coffee off the counter when no one's looking. Regardless of my irritation, if I'm going to consume the coffee, I pay for it -- or else I can take my business to a different coffee shop or go to city council meetings and try to stop Starbucks from getting approved to set up shop on every corner.

I know a lot of us here have the skills to find workarounds to lots of business models, whether it's paying sub fees or viewing ads or whatever. But a lot of us are also trying to build our own businesses. And at some point, we're going to want our potential customers to play within our business models and do a fair exchange (of cash or ad views or whatever) if they get value out of the goods and services we provide. So it just seems like good karma to not go taking things you don't pay for from other businesses.

As for EE's tactics, a more constructive way to get them to stop is to let them know you don't like it, let others know, and let Google know.

Frequently EE has a solution, just not the best one. In many cases it has put me on the right track.

Oh, and many people don't realize that you don't need to pay: Just scroll to the bottom.

I have to say, the beta of Stack Overflow is looking real good. It has real good potential, I just hope it's not seen as only a .NET resource.

I'd have to see the difference.

I've had some hard times finding error codes and resolution, but in general I'm used to and happy with Google.

You're not selling me with that.

And would they let Google crawl the site? If so, then badda bing, badda boom, we're back to just using Google again.

There have been some great ideas for how this site could be better than a simple search. I'd like to hear the pitch from the founders as to their unique selling factors.

While working on a project about five years back using mainly Microsoft technologies, I discovered that there were a few consultants who had built their business around Windows error messages. Maybe you need to focus on the right segment of this market ;-)

Who hired their service?

http://www.errorkey.com/Search/Index.aspx?q=NoSuchError&...

Ugh, this error message is filled with errors and definitely needs copy editing. My best attempt:

  The search for "NoSuchError" did not match any documents in the system.

  For best results:

    * Type fewer words
    * Make sure you are using the correct spelling
    * Try different word combinations
    * Search the internet (???)

First thought is that this seems to be much the same as http://bug.gd, except without the email thing. This is probably a good thing, if your results are good enough quality.

I like the idea and think it will be powerful once you allow users to post info on how they resolved the errors. But Splunk already does this, no? Will this provide any new / easier / cheaper features?

I don’t think Splunk is an error search engine.

It's a server log aggregation and analysis thing...

I stand corrected. I was thinking of how Splunk lets you search all your own logs for a specific error code. But I see how your use case is different.

You've created a lot of these search things haven't you? What are you planning to do with them all? Any plan on trying to monetize, or are they all just projects for fun?

Our primary goal is to build tools that people can use, and then we will focus more on the money side. By the way the previous engines are already bringing in money; they are fully using the power of affiliate marketing. Check them out: qwili.com, qhoga.com, karatr.com.

Error key is a tool we build for hackers, ourselves it is more for a useful purpose than to make money.

Alas!

http://qwili.com/Search/index.aspx?q=chu-hi&btnS=Search

http://qhoga.com/Search/index.aspx?q=gnomes&btnS=Search

http://karatr.com/Search/Index.aspx?q=replica+rolex&btnS...

I know this is a really new project, and I would love to find a better way to search for errors... but how are you going to avoid "pulling a Cuil"?

The Search For : NullPointerException did not match any document in the system

For best Result:

    * Type less words
    * Type Words Correct Spelling
    * Try Different Word Combinations
    * Search the internet

Please send us your search suggestion to ErrorKey at intelliverb dot com. Tell us what system you want us to research on, let’s build this together thanks.

Is it possible to log all search queries that return 0 (or less than x) results, then report on that every day/week/month?

Would minimize the "I can't find xyz" type of feedback that you may or may not get.

I love that phrase "Pulling a Cuil". Lol.

I'm not sure how it works but 0xC0000005 works and 0x00000000C0000005 does not.

Is the site 64bit friendly?

I think the real solution here is microformats, to be plugged into bug trackers and knowledge bases. Then something like this would be possible (just mocked it up myself):

http://derefr.googlepages.com/google_errors.png

Interesting, try getting exceptions as well, that might be useful for debugging purposes.

entering E_ERROR gives wierd results. If I were you, I'd save all searches, create a static page for that result and index it in google.

Since you're an affiliate marketer...

Good Idea, Thanks. BTW we are not affiliate marketers; we build engines that leverage the power of affiliate marketing to deliver fairly good product search.

What may want to do is this (for all your search properties):

Let's say I am a user and I search for "Will Smith". I receive some results. I then search for "Will Smith Pictures". I receive more results. I then search for "Will Smith Pictures No Shirt". What has happened is that I received results that were not relevant, so I refined or changed my search till I found the item I was looking for.

So, my idea is that when you have a particular session searching, you aggregate the follow up queries for each search, and display them at the top of the search results to keep the users longer on your site and push more ads. Because in most cases, users will not click, but in about 10% of cases, the user will see something that he had not previously wanted to search for, but now wants to search for.

For example, I search for "Angelina Jolie", and as "recommended search" I see "Angelina Jolie affair with Max Klein", I would be curious to learn more about the affair Angelina had with this mysterious and probably dashingly handsome Max Klein, so I basically double my interaction with your site.

Just a quick note: in your credits I think you mean to reference people's Trademarks, not Copyrights (you use the (c) instead of (tm))

Right! We will fix that ASAP. Thanks for the remark.

I like the hints if no results are returned. this may sound more 'natural' though:

    Type fewer words
    Type Words with Correct Spelling

with regards to "Try Different Word Combinations", I assume you mean the ordering of my search keywords ("apple pie" "pie apple")? If so, why does word ordering make a difference?

Honestly I think getting into the search market is a waste of time. Too many fish in the sea, and too many sea monsters.

Also: why no API? :)

API is a great Idea thanks for the hint. : )

Great Idea!!! bookmarked.

Reply on news.ycombinator.com