Abhishek Reddy — Thu, 19 Oct 2006 03:30:42 +0000

I mainly use Konqueror (KHTML engine) these days. Neither the CSS nor XBL schemes work with it. That's the drawback with relying on non-standard features.

It's also highly unlikely to work on any Links variant, Dillo, or wget, all of which I sometimes use. I'm concerned about this because an email address is a fundamentally important piece of information for humans -- not something that degrades gracefully. If, for instance, I'm operating on a remote computer, or a broken/incomplete operating system, these are the tools I'd use to fetch their address. (Been there on several occasions.)

I guess one can justify case-by-case priorities by volume of spam, target audience, and so on. I agree that there is usually a troublesome tradeoff. Yet, while training your spam filters would take more time and effort, it's often better for you (say, as a company) to do that, than for your visitors (say, customers) to encounter difficulties. But yes, sometimes that's just not viable.

It's a good thing you pointed out that Firefox extension too. To some extent, client technology should improve like this to accommodate abuse-preventing measures. I think this will happen in the future, so I tend to reluctantly accept some un-accessible techniques as a stopgap solution.

By the same token, bots may improve to circumvent them... sigh.

Welcome to TWF, by the way. An interesting first thread.

myrolemodel — Thu, 19 Oct 2006 01:07:46 +0000

Abhishek Reddy;208776 wrote: .. right-clicking a mailto link and selecting "Copy email address" to save it to clipboard, instead of clicking it, loading my email client, and extracting it from there.

I think that should still work in Firefox with the XBL method in #1. If you're "extracting it" from the email client to use in gmail or another webmail service there is an extension also for Firefox that opens mailto: links in webmail - https://addons.mozilla.org/firefox/206/

Abhishek Reddy;208776 wrote: Partly the point of a URI is to have machine-readability so that tasks like these can be done.

It's a shame this point has been exploited by automated email harvestors, leaving us with the job of filtering the spam after the fact.
I simply prefer the "prevention is better than a cure" approach, which is why I went for the most transparent method of obfuscation I could find. The CSS trick at least is visible in all browsers. I can always put a form on the same page.. which probably needs captchas these days to prevent direct spam - another example of bad accessibility being the only alternative to spam filtering

Abhishek Reddy — Mon, 16 Oct 2006 11:53:43 +0000

Wow, those are pretty neat tricks. Thanks for sharing.

I would not consider using the first method at all for the general case. It simply supports too few browsers.

I would consider a modified version of the second method. Rather than using tinyurl to rewrite the URL, perhaps mod_rewrite or an Apache redirect in the site's own domain would work:

.htaccess:
Redirect /mail mailto:foo@example.com
<a href="http://example.com/mail" style="text-decoration: none;">(Text-logo of email)</a>

It would not surprise me, however, if harvesters can extract the address from that.

A problem I have with most email address obfuscators is that they disable some operations that I frequently use. For example, right-clicking a mailto link and selecting "Copy email address" to save it to clipboard, instead of clicking it, loading my email client, and extracting it from there. Partly the point of a URI is to have machine-readability so that tasks like these can be done.

I would more readily invest effort in honing my spam filters than obfuscating my mail links.