Hiding your e-mail against spam-robots

Difficulty level:		Information for:
			surfer
"John Doe"layman	expert		web-master

[ Language ] :: [ Main Index ] :: [ Site Navigation ] :: [ Tutorials ] :: [ Downloads ] :: [ Contact ]

The fact is, that presently every third e-mail is spam. It costs about 5 billions (thousands of millions) U.S. dollars per year world-wide. Statistically ordinary user spends over spam 10 minutes every day. All other internet services loose because of spam - users just do not believe, that they are really secure.

One myth is, that if you do not reply on spam, spammer can not know that e-mail address is real and alive. It is not quite truth. If you use Microsoft Outlook, you confirm your address as soon as you open a spam with linked image inside!

As far as I found, there are only three countries all around the world, which presently have legislative applicable to spammers: People's Republic of China, Slovak Republic and Virginia at United States of America. At this situation is not supprisse, that a lot of hopeful projects fails. Maybe the last iniciative of Microsoft's Antispam Technology Group will be breakthrough, at last.

So, the only way how to protect himself against spam, is not to publish your address. Unfortunatelly, on domain it is not acceptable solution.

Tracking spam-robots

There is no other way how to beat something effectivelly, than figure-out how it works.

I have found 8 applications during 3 hours, which can be definned as spam-robots. All of them have abbility to search automatically over internet and/or selected domains and extract e-mail addresses.

Well, I have prepared a few pages and put there a mailbox links in their different form. Normally, the e-mail address is defined at HTML as follows:

Suppose following sentence on page: You can contact me at John.Doe@company.com any time.
Its HTML-source-code is:
You can contact me at <a href="mailto:John.Doe@company.com">John.Doe@company.com</a> any time.

I have try to change individual characteristic components - i.e. string "mailto", character "@", domain type ".com" - into their HTML-codes equivalents, and/or pure text respectivelly - i.e. "at", or "John.Doe at company.com", or "_at_" etc.

Here are the conclussions (of all 8 robots):

All robots found string href="mailto and correctly extract an address.
Six robots found character "@" and extract part between " : " and " " " and/or " < " as address
Two robots found string "at" (no matter what characters are before/after it) and extract part between " : " and " " " and/or " < " as address
No one found "@", resp. separated "at" - e.g. "*at*" when at least one of " : " or " " ", has been coded.
No one found string "at" when no space between word before and after used - i.e. John.DoeatCompany.com.
No one found string "on", even when separated - i.e. John.Doe (on) company.com, e.g.
Two robots found valid domain string - as ".com", ".net" or ".sk" and extract it as address, if " " " or " < " follows.
Two robots found code %40, resp. @ used instead of character "@".
One robot found string "%6D%61%69%6C%74%6F" which is fully coded word "mailto".
No one found an e-mail inside of Adobe Acrobat document.
No one found an e-mails inside JavaScript generated HTML.

Some graphical representation is not necessary. We already know what we need to.

Using HTML codes

We can create an image with our e-mail, of course, we could be sure, that no robot find it, but this is not a point. We can use a full JavaScript to generate a row with e-mail address or even link to stop offline-browsers (see an example below), but this can arise with difficulties also - e.g. "//" means remarks and we should remeber this.

<p>Continue with following page:
<script language=JavaScript>

</script>
<noscript>
~ JavaScript is required ~
</noscript>
</p>

We will stay at HTML codes, because these shows to be a serious barrier, now. We will leave JavaScripts for next generation.

As web-designer, you for sure know, that instead of text - I mean classic alphabet - you can use a codes for rare characters. For example, letter "á" at my name can be written to source-code as á or as á. It will be displayed normally at browser, as you can on the following table (compare with page's source-code):

Char written into source-code:	á	&aacute	&#225
Should be displayed "á" at all columns:	á	á	á

Thanks to small research we know, that only way how to protect e-mail links on web-pages effectivelly is coded them randomly. For sure, we have to code characters "@", " : " and " . ". We shall code string "mailto" - but rather not fully. And we can code address itself, but it turns to be not necessary.

So, our address from beginning should looks like this example:

You can contact me at <a href="ma%69%6Cto:Jo%68n%2EDoe%40compan%79%2E%63%6Fm">
John.Doe@company.com</a>

which will be displayed as following (please see the source-code of this page):

You can contact me at John.Doe@company.com

Please also note, that you will see an e-mail box address translated into human form at status bar.

To help you, I created a small utility, which I have called "E-mail box's Hide".

You can download it from "Downloads" page (together with their source code).

Here is only its picture presented. As you can see, you can code address and reference text at various level. As result you get full <a href tag entry, which you can paste directly to your page source code.

Please note, that if you have any problems, forbid the masking of character ':'. In rare cases there can be accumullated too much. It should help. See also a source code of this page.

You can translate or link this tutorial under conditions written at this domain's "Legal Stuff" page and followed. If any doubts, please contact me.

Copyrights, trademarks and credits are collected here.

Marián Stach
Prešov, Slovakia Central Europe
2003-07-08

[ Language ] :: [ Main Index ] :: [ Site Navigation ] :: [ Tutorials ] :: [ Downloads ] :: [ Contact ]