Re: Spamcop Web Service?

After drinking 3 Pan Galactic Gargle Blasters, Neo Geshel
<[email protected]> mumbled in news:[email protected]: 
> Anonymous wrote:
>> Neo Geshel wrote...
>>> Is there a web service that I can connect to in terms of spam
>>> reporti ng? What about spam submission? If there is nothing
>>> now, will there be on e made available in the future?

A XML-interface would be great!

>>>  I am looking to roll my own implementation of the “report spa
>>> m” pages, to make them much more performance-orientated and cut
>>> out features th at I have no need nor interest for. Additionally,
>>> I would love to implem ent it as an AJAX app on my own server.
>>> The problem is, the only way of reliably doing it without a web
>>> servi ce is via screen-scraping, which breaks every time a
>>> significant change  to the reporting process is implemented. I
>>> would rather not do it via screen scraping, so I am curious if
>>> a web service is or will be available for members. 

It would be great - and it is great! I have build a system which
does excactly what you describe - and yes it does break, when new
features are introduced and changes made - but that actually
doesn't happen that often!

You need to use a module/system which is HTML "object" aware,
which breaks the HTML up in pieces, e.g., the button: text, name,
value, etc.

I've made it as in inline browser in an application, i.e., it looks
like a browser and acts like a browser, but does much more behind
the scene!

>> I am just a user, not an admin, but I would guess that Spamcop
>> wishes to retain full control of what gets reported and what the
>> reporter sees while doing it, and that avoiding false reports is
>> a lot more important than "performance" (which I assume means
>> "report many spams very quickly"). 

You are right. My system could easily be abused to report everything
on sight but used correctly - which I think I do by running it in
either Learn or Unattended mode.

When in Learn mode I click on the checkboxes which need to be
notified when the system doesn't recognize a spam and the system
records the URL's, reporting addresses and my actions.

When in Unattended more it just repeats my actions if the same
spam reappears in my SpamCop reporting queue. Anything which does
not look like something the system has been taught is discarded,
i.e., cancelled.

This of course will not report that many spammails because most
of them differ in one way or another, e.g., a random change in
the URL arguments, but I have implemented regular expression so
I am able to catch any similar URL's. This just works wonders!

>> I went through the same thinking at first, because I get a lot of
>> spam and no (almost no?) ham on one address and would like to
>> bulk-report it, but I came to realize that spamcop has spamtraps
>> that do exactly that without the risk of a false report that I
>> would be causing if I bu lk reported or even sped up my reporting
>> and spent less time eying each spam.  I decided that the best way
>> for me to contribute my effort is to report the very freshest
>> spam that I see, and to take the time to verif y that every
>> report I send is a report of real spam.  I think Spamcop needs
>> quality  a lot more than it needs quantity; they can get more
>> spam any time they wish by setting up more spamtraps.

The "problem" with the spamtraps is that only the abused mail
gateway are reported but not the website URLs.

When I have verified the first spam and tought the system I don't
want to recheck the URLs again and again! I just want to report
it!

> The point is, I do all of my spam “submission” through a
> program called MailWasher. That is the easy part. However, the
> next step is to visit the site and click on the “unreported
> spam saved” link. I catch spam through about 20 separate e-mail
> accounts, which can amount to a rather large amount of spam per
> day.

Same here - several hundred a day - when I am lucky!!!
When I am unlucky there may be several thousands!!!

> What I *don’t* need, is all of the superfluous information and
> extra text boxes that the on-site spam reporting methods provide
> the end user with. All I really need is the two or three lines
> showing the content of the spam (which also asks the user to make
> sure it really is spam), the list of e-mail addys that it will
> report to, and the “submit report” button. I don’t need the
> (often) 200+ lines of detailed spam-parsing info (which still
> comes up, even if you turn it off in the options), nor do I need
> any of the form textareas that allow you to add additional
> comments to the outgoing spam reports. And finally, I do not need
> the summary page that shows me where reports were sent - I would
> rather the page just default to the next piece of unreported spam
> in the queue. 

A XML stream would allow you (and everybody else) to show whatever
you want.

> The only way I know around this is by customizing my own spam
> reporting pages (which would be hosted on my own server), that
> would screen scrape for any “unreported spam saved” links on
> the submission page, and parse those pages for *only* the basic
> information that I need. The easiest way would be via a web
> service making use of SOAP or XMLHTTP. The other option would be
> screen scraping, which would break whenever the layout of the
> critical content that is being searched for changes. 

If only a XML interface was implemented all your (and my) prayers
will be answered! But this will open up for abuse of the system,
so I think this will be in our dreams! :-(

Andrew
-- 
*** The opinions expressed are not necessarily those of my employer. ***
* Software Engineer Andrew Engels Rump * BLIK og ROERarbejderforbundet *
* Immerkaer 42, 2650 Hvidovre * Tlf: +45 3638 3638, Fax: +45 3638 3639 *
Home: N55°41'38.9" E12°29'08.6" (WGS 84) Work: N55°39'50.9" E12°27'47.4"
E-mail: mailto:[email protected] WWW http://www.rump.dk/homepage/andrew/