automated service to detect plagiarism?

Copyscape won't work anymore, I think....

[stapel]

I've been using Copyscape [copyscape.com] to track down infringing copies of my content. My site is free-access, so I haven't tried using their CopySentry service, as this would easily run to a hundred dollars a month. Instead, every two or three months, I've run all my URLs through their interface.

This last time, however, I got an error message saying that only fifty searches a month are allowed per domain name. (Somebody else must have been doing some searches, because I hadn't gotten to fifty pages yet.) And the only informational link provided led to their sales department.

So I'm guessing that Copyscape is now for big or fee-based sites only, and us "little guys" are on our own again. (I sent an enquiry to them a couple of weeks ago, and haven't yet heard back. Draw your own conclusions.)

Is anybody familiar with other similar services? Or are we going to have to go back to doing multiple key-phrase searches in Google for each page?

Thank you.

Eliz.

[hunderdown]

Why check every page? I use that copy-search service too. My assumption--though I would agree that I could be wrong--is that if the most visited pages on my sites, and/or the ones with content that earns the best with AdSense, aren't getting copied, then it's really unlikely that I have a problem with other pages on my site.

Use sites like that to do periodic checks. If you uncover a problem, then it may be worthwhile to pay for a more extensive search.

[stapel]

hunderdown asked: Why check every page?

Because I like all my content...? *wink*

All the pages have ads (one vertical banner in the sidebar), so protecting only the pages with ads would still mean checking all the pages. Besides, I'm frequently surprised by what gets nicked. And something will get nicked.

So my question still stands: Is there anything out there, other than laborious hand searches or paying for CopySentry?

Eliz.

[larryhatch]

Hi Stapel: Sorry I don't know of anything much else. I do hand searches. Yes its kind of tedious,
but suspect I do a better job (knowing my turf). I don't have to do it very often any more. -Larry

[hunderdown]

stapel, I like all my content too. I just don't want to spend more time on this than is necessary. I think my method will detect major copying that could actually hurt my site. If I miss a page on a site that isn't getting any visitors anyway, I don't think I need worry about that.

[stapel]

hunderdown: If I miss a page on a site that isn't getting any visitors anyway, I don't think I need worry about that.

If a site gets no visitors, that would likely include Google, so I wouldn't know about it anyway. *wink*

My site is fairly specialized, and those doing the plagiarising are usually directly undermining my "market". These other sites might not get many visitors, but those visitors would otherwise have come to my site.

I've even had some of those visitors contact me directly, wanting to know why they're being charged for access to something that I'm providing for free on my own site. So having the copies out there hurts my "market" and offends users. It is, we've concluded, "worth it" to chase down the plagiarism.

Eliz.

[Beagle]

[Deleted by owner - Because I realized I was misinterpreting what I thought was a misinterpretation, but now I'm not sure what I was understanding or misunderstanding, if anything, so have decided to keep my mouth shut ;-) But as far as I can tell, there's no way to actually delete a post.]

[oddsod]

String from OP [search.yahoo.com]. Doesn't seem to have been copied yet.

What's wrong with trying strings of text in your favourite SE? Why do you say "multiple key-phrase searches"? You very concerned even about fair use reproductions?

Can't you identify a core sentence (or sentences) in each page and create a db? Then automate the searches?

(I used a decent SE in my example above because Google hasn't indexed this page yet ;))

[hunderdown]

stapel, thanks, sounds like you face a very specific problem, and what works for me obviously doesn't for you. Good luck!

[stapel]

oddsod asked: What's wrong with trying strings of text in your favourite SE?

With some four hundred pages, checking multiple strings in each becomes quite time-consuming.

oddsod asked: You very concerned even about fair use reproductions?

What do you mean by "fair use reproductions"?

We haven't had any cases which might fall under the "fair use" exception. By definition, "fair use" of my educational materials requires that the use be immediate (short term, rather than a permanent part of the course), limited (to the classroom, rather than web-viewable by the world), transformative (not using an algebra lesson as an algebra lesson), with no means of replacement (but we do offer site-licensure, so they don't "have" to plagiarise), and not negatively affecting marketability (such as taking for free what we offer for a fee).

oddsod asked: Can't you identify a core sentence (or sentences) in each page and create a db? Then automate the searches?

We've tried that, with disappointing results. The results for some pages omitted "hits" that we knew about, and other pages returned enormous lists of irrelevant results. Plus the search engines have conditions of use which discourage or prevent this sort of use of their service.

A representative of Copyscape has contacted me about a new program they're thinking of trying. I will post a follow-up if/when I have useful information.

Thank you.

Eliz.

[stapel]

Update: You can now buy a five-cent-a-search plan (pre-paid), which falls somewhere between the free interface and the CopySentry plan. You still have to enter the URLs by hand (like the free interface), but you can have as many searches as you like (rather than the 100-URL max with the CopySentry plan). And the fee is less than the twenty-five-cents that CopySentry costs per URL.

Advisory: In this new plan, a "search" is anything that you do through the Copyscape interface. So suppose you enter "mydomain.com/page1.htm" into the interface, and Copyscape returns five "hits". Four of them look like they are copies of your page. So you click on the Copyscape links to view the hits with the copied text highlighted. Those four views also count as "searches". So the one URL would burn off five of the searches you'd pre-paid.

Eliz.