For me, there is nothing worse than hurrying to find something on Google, clicking a search result and then finding that the site you were looking for requires registration. You enter your details into the registration form and then have to wait for 5 to 10 minutes before you can access the site. It seems like a very time consuming method for a person to just read one web page.

Bogus data

One of the reasons for site registration is to get a better idea on the sites audience. The stupid thing is that many people like myself have and never will give the correct data to companies that we do not trust. When I do have to register for a site, I will just enter bad data. 

When the webmaster uses this data, his error rate must be so high that any deductions that are made from the data have a high likely hood of being wrong. 

Extra password

Another thing that a registration form gives me is yet another username and password combination. Someone would have to be very stupid to use the same username and password for every site that they use so I have many names and passwords. The problem with this is that when I forget what my username and password is and do not know which email account I used to create the account, it is quicker to create a new account that try and contact the webmaster. This means that on many sites, I am likely to have at least 2 or more accounts.

Humans only

One thing that registration forms are used for are to verify that a visitor is a human and not a program. This should not be the sole reason for using a registration form as humans can be verified through the use of a captch image verification instead. 

Less visitors

One side effect that sites may notice when using a registration form is that less visitors continue to use their site and they have a higher percentage of visitors that do not return. This is because most people see registration forms as a waste of time when they can get the same service from another site that does not have a registration form.

Bug Me Not

The registration form problem has been around for a while and it used to be solved by a site called Bug Me Not. This site described itself with the following:

You're browsing the web and you click a link to an article on a site (let's say nytimes.com) but instead of getting the article you get a screen asking you to login or register. Infuriated at the idea of pointlessly registering for yet another site you turn to your good buddy bugmenot.com

It also gave the following instructions on how to use the service:

STEP 1: Make a note of the website address your are trying to access. For example:
http://www.nytimes.com/cnet/CNET_2100-1024_3-5567274.html
Or even just:
www.nytimes.com
STEP 2: Visit http://bugmenot.com
STEP 3: Enter the address from step 1 into the box and press the "Get Logins" button
STEP 4: You should now be presented with at least one username and password. Make a note of them.
STEP 5: Go back to the site you were originally trying to access in step 1 and proceed to login with the username and password you noted in the previous step.
With any luck you should be able to access your article now!

Bug Me Not eventually succumbed to pressure. It now allows website to request that logins are not stored for their sites, and is no longer a solution.