The most important feature I see is that it's automatic. This will save a lot of time, which for businesses is rapidly worthwhile. Even when you can have a link checker running in the background on your own machine, it's still very distracting.
Comparing it with the one or two automatic link checkers in existence, my goal is quite different from them. They have stagnated and are designed around HTTP status codes. Unfortunately, that's not all there is to the web. Links can (and far too often do) break completely without returning 404—they may redirect to a 200 OK "not found" or "under construction" page; there's more: links can be broken because they go to different content. Detecting changes like that is also valuable and is one of my many longer-term goals.
Another idea that I have which I haven't seen anyone else using is link hash checking. Getting sent to the wrong part of a long page is often as bad as a 404 Not Found.
Finally, my goal is to ensure that if a link ever existed, it shouldn't be broken: someone else might be linking to it. None of the competitors have anything like that, either, focusing on the current state of your site.
Those are just a few of the things that I have been working on with it.
Certain patterns would certainly not be able to be checked; most notably, any #! links would be blacklisted from checking. It'll be a learning process, improved over time. Hopefully it'll get to the stage where adding machine learning to it will be worth the effort. But for the rest of cases, it's quite straightforward: for any hash, check that an element of that ID exists in the document. Certainly there will be false errors reported with this, but there will be a very significant number of genuine problems found by it.
I'm not quite sure what you mean by this; are you suggesting that the packaging for it is wrong, or that the it would not be commercialisable, or that I haven't done a good job of explaining it, or something else? If this is a feature, what would the product be?
I have been intending this as a commercial product. What would you do differently? (I'd like to learn more so I can get it right or if it really is fundamentally wrong work on something else instead—I still have a few other ideas of things that I want to do.)
However, as a commercial product, I don't see it as successful. Reason why is nobody stays up at night worrying about whether their links still work. (don't get me wrong, it's BAD if they don't work. but it's not what they're worry about at night). They worry if their site goes down for a long period of time. They worry about DDOs attacks. They worry about search engine penalties. But rarely do they worry about links not working. Why does this matter? Because that means people will rarely seek for your product. You will need to do a lot of convincing/educating potential users.
Another example of validating the idea/market before building the product. You don't even need to register a domain name. No CPC cost, just post on HN.
Chris, how much email sign ups are you looking for to go forward with the idea?
As a matter of fact, I have done a little building of the product and figuring out of other logistics such as payments (not so easy in Australia at the time I was planning it—the past couple of months have been marvellous for us with Pin and Braintree both dropping their monthly charges and Pin very easy to get started with), before I came to my senses and realised I should be validating it instead first.
Frankly, my intent is to go ahead with it in some degree as a side-project whether I have an audience or not, because there are various things in it that I know I want that I haven't seen anywhere else in the couple of years that I've been vaguely planning it. A few enthusiastic people interested in it (and willing to pay for it) will change that so that I will probably make it the main project I work on, potentially skipping employment (I'm just about to finish my Bachelor of Software Engineering degree) while I try it out, instead of a side project which will take a few years to get to market if market there be.
I'll be trying a different tack of HN posting tomorrow with a blog post on a related subject—I'll see how that goes too.
9 comments
[ 2.5 ms ] story [ 26.6 ms ] threadComparing it with the one or two automatic link checkers in existence, my goal is quite different from them. They have stagnated and are designed around HTTP status codes. Unfortunately, that's not all there is to the web. Links can (and far too often do) break completely without returning 404—they may redirect to a 200 OK "not found" or "under construction" page; there's more: links can be broken because they go to different content. Detecting changes like that is also valuable and is one of my many longer-term goals.
Another idea that I have which I haven't seen anyone else using is link hash checking. Getting sent to the wrong part of a long page is often as bad as a 404 Not Found.
Finally, my goal is to ensure that if a link ever existed, it shouldn't be broken: someone else might be linking to it. None of the competitors have anything like that, either, focusing on the current state of your site.
Those are just a few of the things that I have been working on with it.
I have been intending this as a commercial product. What would you do differently? (I'd like to learn more so I can get it right or if it really is fundamentally wrong work on something else instead—I still have a few other ideas of things that I want to do.)
However, as a commercial product, I don't see it as successful. Reason why is nobody stays up at night worrying about whether their links still work. (don't get me wrong, it's BAD if they don't work. but it's not what they're worry about at night). They worry if their site goes down for a long period of time. They worry about DDOs attacks. They worry about search engine penalties. But rarely do they worry about links not working. Why does this matter? Because that means people will rarely seek for your product. You will need to do a lot of convincing/educating potential users.
Chris, how much email sign ups are you looking for to go forward with the idea?
Frankly, my intent is to go ahead with it in some degree as a side-project whether I have an audience or not, because there are various things in it that I know I want that I haven't seen anywhere else in the couple of years that I've been vaguely planning it. A few enthusiastic people interested in it (and willing to pay for it) will change that so that I will probably make it the main project I work on, potentially skipping employment (I'm just about to finish my Bachelor of Software Engineering degree) while I try it out, instead of a side project which will take a few years to get to market if market there be.
I'll be trying a different tack of HN posting tomorrow with a blog post on a related subject—I'll see how that goes too.