Prevent search engines from showing outdated docs

[hamogu]

Current marx documentation is a the canonical address space.mit.edu/cxc/marx but all the old version are still available at marx-5.0, marx-5.1 etc. I want users who come by google to automatically find the new version. Two options would be:

• add the old path to http://space.mit.edu/robots.txt and update after every release.
• modify the old webpages to include a box saying "This is an old version. Please see..." In principle, I prefer this method, but it requires me to update the old docs again after uploading the new ones. It would be kind of annoying to rebuild the old docs with added content, so I'd better add a switch when building them in the first place (like "include this file", where "file" is originally empty and added by me later. Second, it would be useful to point to the new canonical version of the same page, not just the top level marx. However that requires individual updates on every page and what to do if the new pages does not exist? Maybe I find some javascript code somewhere that activates a box only if a newer file is present and constructs the link if possible. In any case, I would still have to go back to change the old docs that are already on the server (or go back to the robots.txt method for those).

[hamogu]

A tried a few searches. The very old pdfs for marxs 3.0 and 4.0 come up, but all the links I've seen to web pages point to /marx/ with no version number attached to this is not a problem right now. I should just think about steps to prevent this from becoming a problem in the future (CIAO needs to address that now).

[hamogu]

https://support.google.com/webmasters/answer/139066?hl=en#2 could be used to make sure the canonical page in indexed not the specific page.

[hamogu]

Here are some notes from a search to find the appropriate Javascript commands to insert something if present: Insert something:

var text = document.createTextNode('the text');
var child = document.getElementById('childDiv');
child.parentNode.insertBefore(text, child);

check if present:

$.ajax({
    url:'http://www.example.com/somefile.ext',
    type:'HEAD',
    error: function()
    {
        //file not exists
    },
    success: function()
    {
        //file exists
    }
});

This is even easier: Client Side Includes from http://ask-leo.com/how_do_i_include_one_html_file_inside_another.html

As I mentioned above, there is no actual syntax for a client-side include but we can mimic one using Javascript. For example:

<script src="b.js" type="text/javascript"> </script> 

When encountered, the browser downloads the script "b.js", executes it, and prints any output that the script might generate as if it were inline HTML. Technically that's not an include but the script "b.js" could be nothing more than a series of javascript "print" statements such as these:

document.write("<table>")
document.write("<tr>")
... and so on 

"Welcome to 'include' files - an incredibly powerful facility that can do this, and so much more, on your web site."

You can see it's "printing" the HTML you want included. In other words, if you can format your include file as a series of javascript prints, you can use client-side include to insert it.