Hacker News Reader

Comment by mountainb | Hacker News Reader

Link rot has convinced me that the web is not good for its ostensible purpose. I used to roll my eyes reading how academic researchers and librarians would discourage using webpages as resources. Many years later, it's obvious that the web is pretty bad for anything that isn't ephemeral.

View on HN

We have deposit libraries in the U.K., such as The British library and Oxford University's Bodleian. When you publish a book in the U.K. you are supposed to offer a copy to these institutions.

If we had legal deposit web archiving institutions, then academics, and others, could create an archive snapshot of some resource and then reference the URI to that (either with or without the original URI), so as to ensure permanence.

You can also copyright databases in the US as well as web content. It is similar here. However, many people do not actually register their copyrights to their web content. There is also a lot of misinformation about automatic copyright in the US as it relates to web content. While there is 'automatic copyright,' due to recent court decisions, you must have a registered (paid for, deposited with the copyright office) copyright in order to file a lawsuit for copyright infringement.

I agree that this is what should be done more often for durable content. This is another reason why social media sites are bad: the users technically have automatic copyright to their contributions, but they often also agree to grant licenses to the content website for things like their photos. The US Copyright Office is releasing a new application format intended for blog posts and other short form content that will hopefully be more straightforward to use.

>I used to roll my eyes reading how academic researchers and librarians would discourage using webpages as resources.

While this is true in general, I am amused that this is not true for citing wikipedia. Wikipedia can be trusted to remain online for many more years to come. And it has a built-in wayback machine in the form of Revision History.

Usually a reference indicates that an author believes something to be true, but won't explicitly state their reasons. It isn't just a statement of where information comes from, but a justification for trusting that information. If the reference is from a reputable source, then it indicates that this belief is justified. If an author believes something to be true because they read it on wikipedia, then that belief probably isn't justified, because the reliability of wikipedia content is mixed.

Good quality information on wikipedia often refers back to published sources, and at the very least an author should check that source and refer to it, rather than wikipedia itself.

Try following the references on big Wiki pages and you will see why Wikipedia pages are nightmarish for any kind of research. This is important when you are trying to drill down to the sources of various claims. Many major pages relating to significant events and concepts are riddled with rotted links.

The page can be completely correct and accurate, but if you cannot trace the references then it cannot be verified and you cannot make the claims in a new work as a result. The whole point of references is to make it so that the claims can be independently verified. Even when there isn't a link rot problem you will often find junk references that cannot be verified.

Wikipedia isn't a bad starting point and sometimes you can find good references. But it is not anywhere close to reliable: just trace the references in the next 20 Wiki articles you read and your faith will be shaken.

After someone published an authoritative ftp listening, so many people panicked as their were out of date and insecure versions so rather than patch they all went dark.

Anyone doing research just got screwed.

So many papers have code listed to places that don’t exist anymore.

Reply on news.ycombinator.com