With a search engine marketshare of less than 10%, Yahoo! is already largely seen as irrelevant when it comes to general search, especially since it began outsourcing its search results to Bing!

However, there is at least one area where Yahoo! has remained a critical player: Plagiarism detection.

Simply put, many of the most popular plagiarism detection services take advantage of Yahoo! Search Boss API (Application Programming Interface), which has made creating a plagiarism detection service both affordable and relatively simple.

So, as Yahoo!’s future hangs into balance, so does the future and capability of many of the Web’s best-known plagiarism detection services including Copyscape, Plagium, PlagScan and PlagAware, all of which use Yahoo! either exclusively or in part to find their results.

To be clear, there’s no immediate threat to Yahoo! BOSS and its closure has not even been mentioned. This is purely an academic exercise.

However, with such uncertain times ahead for Yahoo!, the question gets raised, what would a Yahoo!-less plagiarism detection landscape look like? The answer isn’t very clear.

Why Yahoo! is Important

Without using a search API of some sort, a plagiarism detection service would have to crawl websites and create its own index, a time-consuming and expensive process that would cause the services to be prohibitively expensive. Fortunately, most major search engines offer APIs that enable plagiarism checkers, as well as other services, to tap into their indexes for a relatively easily and cheaply.

Many plagiarism detection services began using Yahoo! BOSS over competing offerings for a simple reason: Cost.

Historically, Yahoo! BOSS was a free service. But, even after Yahoo! began to charge for the service (shortly after it began to use Bing for search results) the cost of using Yahoo! BOSS was still many times cheaper than using Google’s Search API.

This is why many of the best-known plagiarism detection services are built either in whole or in part on Yahoo! BOSS.

This cost is important because performing a single plagiarism check, usually, requires multiple API queries (the exact amount depends on how the service handles queries, the length of the work involved and other factors). As such, these API costs often become a major expense for these services.

But more than just a cost issue, the presence of a competing API to Google also offers a different perspective. Being able to tap multiple indexes of the Web rather than just one has the potential to ensure the maximum number of results are returned, especially since the different indexes often catch different content.

In short, without the Yahoo! Boss API, we are likely looking at a much more expensive and more limited future for plagiarism detection.

What Does a Yahoo!-less Future Look Like?

If Yahoo! BOSS were to go away, the future is definitely a difficult one for many plagiarism detection services.

Some, such as Copyscape and Plagium, already mix results from multiple sources (Google/Yahoo! and Yahoo!/Bing respectively) and would likely just lose some of their fidelity in their results. Copyscape would, arguably, be in a better position than most as it began life using the Google API.

Others, such as PlagAware and PlagScan, both of which use (or seem to use) Yahoo! exclusively would be forced to write a completely new backend for their service. This could have a drastic impact on how they detect duplicate content and how effective they are (better or worse).

Higher-end services, like Attributor, which use their own index of the Web would be unaffected by any change or closure of Yahoo! BOSS and may even have their position strengthened.

All in all though, there would be a major shuffle ahead for plagiarism detection services as they looked to fill the void left by Yahoo! BOSS.

Where Would the Refugees Go?

Those who depend on Yahoo! BOSS, if they wanted to stay open, would have a tough choice ahead of them as there are only two (major) providers who would remain.

1. Google: Google’s API is definitely robust, as is Google’s index, but is also much more expensive than Yahoo! BOSS.
2. Bing: Bing’s API is much less established and not as well regarded as Google’s but it is free for unlimited queries, just as Yahoo! BOSS was. However, the API’s TOU may pose challenges in some cases, specifically related to advertising requirements.

In short, for developers it’s a choice between an established and robust API that is more expensive and a newer one that comes with limitations on how the results can be used.

Most would likely go with Bing as it is the most natural replacement (especially since Yahoo! results come from Bing) but it remains to be seen if Bing’s results can compare with Yahoo! or Google’s for this purpose.

That would be something very interesting to test in the future.

Bottom Line

To reiterate the good news, there’s no immediate threat to Yahoo! BOSS at this time so all of the above is merely hypothetical. There has been no talk of closing Yahoo! BOSS and, given that it currently is a revenue generator for Yahoo!, it isn’t likely to be first on the chopping block.

That being said, the turmoil at Yahoo should give cause for concern to those who rely on the Yahoo! BOSS and the time may well be now to start looking at alternatives.

After all, if the end of Yahoo! BOSS does come, it will likely be sudden and it may be difficult for companies that rely on it to quickly reconfigure their products.

Even if it seems unnecessary at this time, preparing for the possibility may be the best move these services can make.

Have any suggestions for the 3 Count? Let me know via Twitter @plagiarismtoday.

1: Amazon, Dropbox, Google and You Win in Cloud-Music Copyright Decision

First off today, music locker service MP3Tunes, though found liable for not complying with DMCA notices, has been handed a major win by a District Court and may be paving the way for Amazon, Google and others to launch more efficient cloud music services. In the ruling, the judge gave his approval on both the fingerprinting of audio tracks, which means that MP3Tunes needs only store one copy of a previously uploaded file, and the sideloading of audio tracks, meaning that users can load tracks from elsewhere on the Web. The ruling is a boost for Dropbox, which uses the same fingerprinting technology to save on file space and bandwidth, as well as other music locker services, such as Amazon and Google, that may want to use these features. However, EMI and the other publishers involved in the suit likely plan to appeal the ruling.

2: Oracle Wants Copyright Claims to Remain in Google Suit

Next up today, in its lawsuit against Google over the company’s use of JAVA in the Android operating system, Oracle has filed a objection to keep a copyright element of the suit intact. Google had previously sought dismissal of some of Oracle’s claims stating that API elements were not copyrightable. However, Oracle, in its objection, has asked for that to be resolved at trial saying that no court has ruled as such and that Google requires its partners to respect copyright in its own APIs. Google will have a chance to respond to the objection.

3: It’s Not Over Yet, Jammie Thomas: RIAA Appeals Damage Reduction

Finally today, in a widely expected move, the RIAA has appealed the most recent ruling in its case against alleged file sharer Jammie Thomas-Rasset. The lower court last month reduced the damages she would be ordered to pay from from $1.5 million to $54,000 however, the RIAA is specifically appealing whether Thomas-Rasset’s alleged sharing is “distribution” under the copyright act, which the lower court ruled it was not. This could set the stage for a third trial in the case, likely on the issue of damages alone (for the second time).

Suggestions

That’s it for the three count today. We will be back tomorrow with three more copyright links. If you have a link that you want to suggest a link for the column or have any proposals to make it better. Feel free to leave a comment or send me an email. I hope to hear from you.

Want the Full Story?

Tune in every Wednesday evening at 6 PM ET for the live recording of the Copyright 2.0 Show or wait and get the edited version Friday right here on Plagiarism Today.

The 3 Count Logo was created by Justin Goff and is licensed under a Creative Commons Attribution License.

Copyright non-repudiation service Myows announced an API challenge for developers last week. The idea is that the developer who generates the best plug-in, add-on or standalone application for their API, meeting the criteria in the announcement, will a prize, namely an Apple iPad.

The runner up will also receive an iPad and all of the applications will also be put before a jury of “industry heavyweights” to award a jury’s prize of $1,000.

Specifically, any submitted application or plugin must do the following three things:

1. Use the new Myows API to develop a plug-in, add-on or stand alone application that assists other Myows members in registering, protecting and/or managing their rights
2. Develop this solution in a way that allow users to access key functionality in 4 steps or less
3. Implement your code and design with integrity, simplicity and elegance

If you are interested in competing, Nettuts+ has two videos to help you along. The first shows you how to register a new application and the second how to build a very basic app.

For those interested in competing, the deadline is May 15. One can submit an application by posting to the API category in the Myows forum. Public voting will take place between the 15th and the 31st and the jury award sometime after that.

I look forward to seeing what interesting applications people come up with for this challenge. Personally, I am hoping for a good WordPress plugin.

Disclosure: I am a paid consultant for Myows.

Though Twitter plagiarism has never raised the copyright questions that come with blog plagiarism, it certainly has caused some emotional outbursts. For some time now a safe way to start a fight among Twitter users has been to ask about best practices when retweeting. Though everyone seems to agree that attribution is necessary, how to provide it has been a tricky matter.

Some provide it at the beginning of the tweet (RT @username TWEET), others at the end (TWEET via @username) but other than the RT as the standard abbrevation for “Retweet”, not much has become consistent.

This becomes worse when a tweet is too long to retweet properly, meaning the original tweet plus the attribution is too long to send out. This forces a devil’s choice of either changing the original tweet, thus possibly changing the meaning on accident, or lobbing off attribution.

Help may be on the way though, yesterday, Twitter announced that they were making changes to both their user interface and their API to make retweeting easier and more consistent.

This could finally put an end to the problems with retweeting and help separate the malicious plagiarists from those who simply haven’t learned how to RT properly.

Retweet Changes

The basic idea of the new change is that, instead of the retweet appearing to be from the person doing the retweeting, it will instead appear to be a tweet from the original author, with a line below it (similar to the “in reply to” line that exists now) indicating who sent out the retweet. See draft mockup below:

For example, if I retweet something by my co-host @ifroggy (Patrick O’Keefe), even if you do not follow his account, you’ll see his icon and the original tweet appear in your timeline. Underneath the tweet will be a short line indicating that it was retweeted by me.

This ensures that the content of the tweet will be easily identified as being from Patrick and will introduce him to the my followers in a much better way than a simple @mention in a retweet. The solution also makes it much cleaner for people to retweet other’s retweets, avoiding the “RT @username RT @username” kludge that has existed for some time.

Developers should be able to work this into their existing Twitter applications and the changes should be available to all users some time within the next few weeks.

Tweet This

The new system, if it goes through as planned, will solve nearly all existing problems with retweeting. Tweets can’t be too long to retweet since the original one is used, there’s no need to argue over how to properly attribute a tweet as there will now be a formal system in place and it makes it easier to spot the spammers and other abusers as those who don’t use the system to retweet other’s work will likely have more selfish intentions than simply be ignorant of the rules.

However, the system isn’t without its detractors. Commenters to a recent post on the changes in TechCrunch expressed concern that they won’t be able to edit the Retweets they send out, making it impossible for them to add their own element to them.

Indeed, the process does appear to be fairly automatic. Once you click the link to send the retweet and confirm you want to send it, the tweet goes out without any opportunity to edit. However, this also solves the problem of fake retweets being sent out using celebrities names, which became a major problem for a while in July.

The question is whether it is worth the ability to edit retweets for the standardized system, better exposure to the original author and convenience of the impending new system.

The answer to that will likely depend upon one’s own Twitter habits.

Bottom line

The new system will hopefully clear up and fix a lot of the problems with retweets that have been plaguing Twitter. It’s easier to use, provides better attribution and creates a good standard for use. However, it clearly doesn’t fit all cases or all uses of the existing RT system.

As such, it seems likely that the existing methods for retweeting will hang around, at least for limited use, and that new standards will be developed by the community that will address those limited cases.

However, if those standards are anything like existing ones, they won’t be very consistent and may lead to some inadvertent strife. Still, a partial solution that covers most retweet cases is still better than no solution at all.

Goldstein took special issue with a series of recent cases where copyright licenses were being ignored, by users of the Flickr API, the latest of which involved making all Flickr images, regardless of license terms, available for download as cell phone wallpapers on the site Myxer (the article mistakenly reports the images as being for sale, though the download, according to comment 40, was free).

It is very clear that many services and companies that have used the Flickr API have violated copyright holder’s rights, either intentionally or accidentally, and that this is an ongoing issue as new services come online almost every day.

So what can be done to fix this problem? What responsibilities does Flickr have in this? The answers, unfortunately, are neither simple nor easy.

The Power of the API

There is little doubt that Flickr’s API is a very powerful tool. It allows third parties to build services and tools that access Flickr and use the images there in new and exciting ways. It is behind many of my personal favorite tools, including Photodropper.

Also, most applications that use the API do so in a way that is fair to the rights of the artists that use Flickr. It is, fortunately, only a small minority that do not. This is because the API makes it simple to interpret the licensing of the images and Flickr’s terms of service for the API requires developers to respect user intellectual property.

However, some have not and those cases pose a great deal of risk to photographers. Since the infringers are using the API, much like an RSS scraper, they have the ability to take almost everything on the site and do with it as they please. This includes, theoretically, selling the works, creating new, high-resolution galleries and using the works in advertising or promotion.

This has many photographers worried and, judging from the comments on the original article, at least some are abandoning Flickr due to these issues.

Flickr’s Role

Flickr, for their part, is in a bad position here. Their powerful API is one of the critical reasons that both developers and users enjoy the site as much as they do. Flickr’s ability to interact with other services has been critical to its success and removing functionality from the API could be very costly to them.

Despite that, Flickr does have both a terms of use that forbids developers from abusing user’s rights and the ability to revoke API keys, thus shutting down services that might be infringing.

However, Flickr has been slow to use this tool against developers, especially those that create products with uses that have legitimate uses. This has not stopped Flickr from shutting down some services trying to access the site, such as it did with the image search engine FeelImage (though FeelImage was not using the API, just a tag search, and has since resumed indexing only CC-licensed material), but such cases usually only take place after a user uproar or if the service is clearly abusive in nature.

The simple truth is that that the vast majority of the responsibility is on users to license their photos correctly and developers to respect those licenses.

Flickr, though it is the middle man, has very little it can do in many cases.

What Flickr Should do

This is not to say that the site is immune from all responsibility or criticism in this matter. There are several things the site can and should do to reduce the number of such incidents.

If I were to make suggestions, I would include the following:

1. Clearer User Licensing Terms: The image above and to the right is what I see when I log into Flickr’s privacy options. The options are confusing and overlapping. “Sharing” a photo, for example, allows users to embed or “blog” a photo, which is yet another option, there is also no clear way to remove an image from the API (Note: You have to disable public searching on “3rd Party Sites”) and it is unclear how any of this meshes with Creative Commons Licensing. If this is confusing to me, I can imagine many users feel overwhelmed.


2. Quicker Disabling of API Keys: If a developer is infringing on the rights of Flickr users, their API key needs to be disabled, at least until a fix can be made to their system. Though Flickr is understandably uneasy about banning developers for a coding mistake, they could allow such sites 72 hours to correct the problem before disabling them.


3. Licensing Trumps API Permissions: Under the current system, the API setting in Flickr trump the licensing settings on the photograph, it either should work the other way around or the user should be given the option to decide which is more important. Otherwise, the copyright licensing is fairly meaningless.

As always, I am seeking other suggestions as to what Flickr can do so please feel free to add your ideas in the comments below. Since I am not a heavy Flickr user, I realize my input is limited.

Conclusions

In the end, the responsibility to respect licenses will always fall on the developers. As with any API, the developer will have the ability to disregard both the terms of use and the rule of law, but have a duty to respect both their own agreements and the user wishes.

While there are steps that Flickr can and should take to reduce this problem, the issue of Flickr-based tools ignoring licensing terms falls squarely on the shoulders of the developers that made them.

If developers do not bear responsibility, legally and ethically, for the works they create, then there is absolutely nothing to stop them from abusing the system even more. They, as well as the users who abuse the tools they create (in some cases), need to be held accountable first and foremost.

Though the frustration with Flickr is understandable and certainly is some grounds for it, they are not the ones who abused the system nor are they the ones who made the mistake.

They just paved the road for those who did.

Though I’ve been very critical of Copyscape in the past on this site, it is clear that they are the only company currently serving this niche and many people are excited about their services.

However, the biggest criticism they’ve faced is that their free service is too limited (only ten results per search, a few searches per month) and that their main paid service, Copysentry, is too expensive.

To answer those criticisms, Copyscape recently began offering a series of premium services, targeted in the space between their free and automatic offerings. The idea to be affordable enough that more bloggers and Webmasters would upgrade to their paid offerings and not rely so heavily on the free service.

Unfortunately, as great as the idea is, the execution of this service severely limits its effectiveness, thus only prolonging the agonizing wait until a real plagiarism detection/cessation service comes to light.

What You Get

At the core of Copyscape Premium is an unlimited version of its free service. For a rate of five cents per search (minimum $5 purchase required), you get unlimited searches with unlimited results. You also get the ability to paste text directly into a text box for searching, instead of just submitting Web pages, and an API that you can use to automate and refine submissions to the service.

However, the service goes above and beyond that as well. It offers a plagiarism case tracking service that lets you add discovered cases to an inbox, add notes to them and close them out when they’re handled. Users also have the ability to blacklist certain sites from the results, useful if your own site keeps popping up, and, according to the site, a more robust detection algorithm.

On the surface, it sounds like a great deal. Five cents per search is very reasonable and the case tracking technology alone, if done right, could be worth the price of admission by itself.

That’s why it is so frustrating, and heartbreaking, to report that the service itself is a monumental letdown in its current form. Perhaps not much in the way of a waste of money, but a tremendous waste of time.

My Experience

The process of setting up an account with Copyscape Premium was simple enough. I registered my username and deposited $5 into my account for a total of 100 credits.

I then sought out one of my earlier poems, one that I knew was plagiarized and copied heavily, and searched for it using the URL. Copyscape spun on it for a few seconds and returned me to the results page, which had only one sentence on it “No results found…” (see screenshot above).

Confused, I figured Copyscape was struggling with the large number of comments on my URL. I copied and pasted the text of the poem into the text search option but it was to the same result (see second screenshot).

I then exited Copyscape, performed a simple Google search for the poem and instantly found dozens of suspect pages. Copyscape was simply not picking up any of the copies of the work.

I repeated the experiment with four or five other works, using a combination of URL and text searches. After ten searches, Copyscape failed to find even one plagiarist, even though all of the works I searched for were confirmed to have plagiarized copies on the Web.

Copyscape did return a single result on at least one of the text searches, but the result lead to a legitimate use of my work, not a plagiarist. It would be a theme that would repeat itself over and over in my attempts to use Copyscape Premium.

But Wait… There’s More

Adding to the ineffectiveness of their search is the fact that Copyscape does not allow you to add text searches into your “View Cases” inbox. As a result, I could not test their tracking system with the results I had on hand.

To do that, I had to punch in the URL for Plagiarism Today. That turned up 17 results but almost all of them were junk. The results included people quoting from my sidebar, sites using similar plugins (such as the Share This! plugin) or sites legitimately citing articles. The only one with a substantial amount of reproduced content was a Numly Verification Page, which I created myself through my Numly plugin. Still, it was good enough to test their case tracking service with for the moment.

Unfortunately, that service left a great deal to be desired as well. You create a case by clicking on a button in the Copyscape frame that’s visible when you visit a similar site. Once you create a case, you can add notes to it, close the case and reopen it. However, there isn’t much else you can do. Closed cases remain in the inbox, there are no sorting tools, and the notes are not lengthy enough to add any correspondence or information beyond a few words.

A simple spreadsheet or, even better, a full database could serve this purpose significantly better. It is convinient, but ultimately useless.

All in all, I spent about 20 of my 100 searches. I discovered no actual plagiarists and the vast majority of my results came back with no matches found. That is especially worrisome since my text searches should have, in theory, turned up my own site at the very least. After all, I only blacklisted my own domain for the last few searches.

Fortunately, the process only cost me $5 ($1 in actual searches) but it did eat up over one hour’s worth of time spread across two days. In that time, I could have easily set up dozens of Google Alerts and received automated, free and effective plagiarism detection.

Conclusions

In the end, there is very little good that I can say about Copyscape Premium. The results were ineffective, the tracking system was almost completely useless and the entire service seemed inept at finding plagiarists even when they were hidden in plain sight.

I really wanted to enjoy and rave about this system. I wanted it to be the great comeback for Copyscape in my eyes. Sadly, it was nothing of the sort.

The price point is great, the searches are lightning quick and the idea itself is wonderful. If Copyscape Premium actually did what it promised to do, it would easily be the best solution available right now.

However, in its current state, I can not recommend it at all. It is nothing more than a waste of money and time.

Hopefully, one of the companies that is developing plagiarism-detection products will step up and produce a real service that will fill this obvious void. The technology seems to exist, we’re just waiting for someone to put together all of the pieces.