Technorati authority dropping due to anti-spam initiatives?!

29 10 2008

In the previous post (Technorati rank & authority dropping like the stock market) I mentioned the acute overnight drop of my Technorati Authority from ~46 to 5 and the loss of many backlinks.

Here I suggested that this dropping in authority might be connected to the loss of Google backlinks.

From the Technorati discussion forum it is apparent that many other bloggers are having similar problems: the loss of blog reactions and thus “authority”.

Going through some of those discussion, I found that answers of the administrator gave a clue to the cause of the vanishing pings.

In the discussion string http://support.technorati.com/discussions/topic/4655 the administrator wrote on October 17, 22 and 28 respectively (see Figure):

[Note the different insight over time and the light hearted tone:

” Hello all, we did a bit of spam cleaning over the weekend…” ]

The last response links to a blogpost of Ian Kallen on October 27, entitled: Data cleanups and mishaps, that clearly confirms that the “mishaps” do relate to (finally) cleaning up Technorati spam in a very rigorous way.

Here is the integral text of the Technorati blogpost.

“Technorati has a number of initiatives in the works to improve the data in our search indexes and analytics systems. Web spam sites (splogs) have long been an issue that we’ve been working to address. The days when pings came only from legitimate blogs are long gone. Including all of the spam and duplicates, Technorati receives over 8 million pings per day. Over 90% are recognized and blocked as soon as they’re received. The remainder is allowed into the system and selectively processed – a large portion is determined to be spam later.

Recently, we’ve been focusing on link farms and pornography sites that have been getting into the system. Link farms are networks of sites linking to each other and other sites with the intention of raising search rankings. Sometimes, these sites link to legitimate blogs to “camouflage” these intentions or simply because the content has been stolen from another site. During a recent scrub of the system, a number of legitimate blogs were misidentified as spam. The flags set on those blogs were reversed, so going forward they are being indexed correctly again. However, some of the link and post data scrubbed from our search and analytics systems could not be reverted. We’re working on upgrades to make that data handling better managed but in the meantime, there are some gaps in certain blog’s data which may affect the authority of blogs they linked to. Additionally, some blogs suffered authority drops due to being the beneficiary of camouflaged links from spam sites being removed (wittingly or not); when those spam sites were removed, so was a portion of the authority of the legitimate blogs they linked to.

We have a number of technology initiatives in the works to improve the scaling characteristics and data quality of our systems. More news will be arriving on that in the weeks and months ahead.

Indeed this explains a lot. As I wrote in previous posts ( Blog Spam and Spam Blogs 1 (see here) and 2 (see here)) many splogs have linked to my blog and much of my content has been and is being stolen by such blogs!!

So I’m punished twice and hard for writing about health related issues (the desired niche for spamblogs selling cialis, viagra and those kind of drugs).

Once by blogs stealing my content and ending up high in ranking (see comment of Wowter and Keith Nockels here) and once by Technorati finally cleaning up those spamming blogs in a rigorous way, dragging me along in their slipstream!

Thanks Technorati! For shooting holes in my ranking, not responding to my mail and not adequately helping those who are hit by your rucksichtloss (excellent German term for what has been done, something like recklessly in English) weeding of the spam blogs that you’ve allowed to exist in Technorati for years! (see this critic in Wikipedia mentioned in my previous post).

Technorati, what are you going to do about it?


Advertisements




Possibly an announcement about possibly related posts

16 05 2008

Related to my previous post “new related reviews in Pubmed” (which appeared possibly an anouncement) is this post on “possible announcement of possibly related post” appearing in WordPress. Are you still with me?

Matt wrote a post “Possibly an anouncement” about the introduction of a new feature in WordPress: Related Posts. In English posts to begin with:

“In a feature we’re calling possibly related posts we’ll now try to show posts related to yours a little section at the end. If we find any posts on your blog that are related, we’ll put those at the very top and in bold. Next we’ll show other posts from around WordPress.com, and finally we’ll check if there’s anything in the mainstream media.”

In the ensuing discussion Matt reveals that “related thingy” as someone calls is not done via tags or keywords, but with all the content in the post. According to Matt “Sphere” creates something they call a “document genome” (LS: Wow! they sequence it!) and use it to do the matching, which can work a lot better than tags and categories”.

It is still in an experimental stage and will be tweaked on basis of feedback and data collected. Introducing related posts may increase traffic between blogs as one can easily go from one interesting item via another to an even more interesting item. A kind of “post-hopping” or lateral navigation, as Matt calls it.

Interesting idea, but is it working?

Well, I surveyed some of my blog posts in a similar way as Hugh from the “I’m not actually a geek” blog to see what the possibly related posts to my posts were. They’re shown below, along with a rating of ‘++‘ for related and interesting, ‘— ?‘ for not related, ‘±’ for posts show a certain similarity. ↔ = reciprocal

A. new-related-reviews-in-pubmed/ (Dutch/English)

  1. From Change the Display Format of Single Citations with My NCBI PubMed® User Preferences (link) ±

(it is about PubMed, o.k. but mature and about the display format, not really about related articles and certainly not about related reviews-which are new ; the site itself is interesting -for me, but not updated for a year (Quite the opposit, Anna updates here blog daily. See somments. Sorry!)

B. Opening UBA? (Dutch)

the first two are from the same site and they link to each other but not to me.
Hé: Dutch posts show the relating feature as well 🙂

C. appropriate-bedside-manners (Dutch/English)

  1. Why paternalism isn’t dead « WhiteCoat Underground ± (about doctor -patient communication, but not about bedside manners or doctors crying
  2. The Art Of Medicine, kind of book review ± (same, and a book review)
  3. Building Physician Practices in Florida through Interactive Websites — ?
  4. At Bedside, Stay Stoic or Display Emotions? ++ (but it is a reference in my blog

The last post (4) is a reference in my blog (yeah, easy!), and does not relate to me, nor do the other ones but 1 and 3 do relate again to 4.

D. etiquette-based-medicine (Dutch/English)

  1. Een nieuw begin (uit de oude doos) — ?
  2. Een al te lange roede — ?
  3. Topartiest — ?

No resemblance whatsoever, except perhaps some english words/an English title next to Dutch ones. The second one is again a nice blog (a dutch biologist, brughagedis!)

Well E, I see at a glance, yields nonsense again. I don’t take the trouble to write this down.

This trial was unsuccessful, indeed. The extremely low yield of NEW interesting posts might relate to the usage of both Dutch and English, but I can’t understand why a new subject as Etiquette based medicine (NEJM-paper) does yield only irrelevant Dutch posts. In addition the related posts do not link back. Till now I never received any hopping guest (except via the WordPress dashboard, news, or tags).

Which brings me to a more generally uttered criticism, that is that Related Posts brings your reader(s) away from your blog, so you loose and don’t gain.

Others criticize is that the posts linked to their blog are not remotely related and sometimes even quite contradictory (vegy food gives meat) or even link to racist and offensive material and that there is no control over the links, even a kind of intrusive to put (unrelated) links below a post without any notification. Somebody noted that It would be great if we had an easy-to-use formatting tool to list related posts of our choosing, not random word-search weirdness from the internets. Others object that tags (added yourselves) already serve this purpose of lateral navigation.

My feelings are quite similar to those of Douglas Bell (commenting on the announcement)

“I’m sorry, but I am astonished at how this feature was added. The particular blog that I run on WordPress.com is a professional blog with important content, and now to see something called “Possibly Related Posts” appearing in the exact same place and format as my own content, displaying links to blogs that I have no control over (and many of these links having absolutely nothing to do with my blog and linking to very inappropriate non-PG, non-English content) is quite infuriating. It’s fine that this feature can be turned off, but shouldn’t it have been opt-in, instead of opt-out?”

Of course there are positive reactions as well, mostly very short comments : awesome, I like this, you guys are great….

But I don’t like it, not when it works this way and I will put the feature off. I might give it a second chance later to see if it performs better.

And the missed traffic to my blog? I don’t mind. I rather have a few interested people here, than just traffic passing by. I hate highways anyway. Prefer the quiet countryside and wait there for somebody to visit me and stay for a while. And of course I pay a visit back.

Note added in proof: some related posts have been refreshed in the meantime.


* The 3D WordPress Logo resembling a yoyo is created by Kate from Nineteen Labs





Delicious Google Toolbar

30 03 2008

Wie nu denkt dat ik het uitgebreid over Del.icio.us ga hebben komt bedrogen uit. Vooralsnog is het bij wat vingeroefeningen gebleven. Ik vond het allemaal niet erg overzichtelijk, maar op de pagina’s van de Afvalchinees en no 33 (Patricia) zag ik wel zeer zinvolle toepassingen. Ik wacht dus nog even met mijn eindoordeel.

Het valt me op dat veel Spoetnikkers als grootste voordeel van Del.icio.us noemen dat Favorieten/Bookmarks/Bladwijzers nu ‘overal’ bereikbaar zijn, zowel op werk, thuis of elders. Als 2e voordeel wordt genoemd dat je verschillende tags aan één bookmark kunt toewijzen en niet langer 1 adres in 1 mapje hebt.

Deze twee functie kunnen ook heel goed door de bladwijzerfunctie van de Google Toolbar uitgevoerd worden (te downloaden zowel in Firefox als IE-versie). Bladwijzers zijn weergegeven als een blauwomrande ster, van binnen wit als de geopende webpagina waar je je op bevindt nog niet gebookmarkt is en geel als dat wel het geval is.

Google bladwijzers zijn heel makkelijk in het gebruik. Feitelijk zijn het ‘tags’, hier ‘labels’ genoemd. Je kunt een adres van verschillende labels voorzien. Een adres komt daarmee dus in verschillende mappen.

Als je labels wilt toekennen, kun je een keuze maken uit alle aanwezige labels (een mogelijkheid die ik niet heb kunnen ontdekken bij del.ici.ous, daar moet je een beginletter invoeren). Ook bij het 2e label worden de keuzemogelijkheden getoond (zie Figuur).

Prachtig dat je overal bij je favorieten kunt. Zeker ook wanneer je computer crasht, hetgeen me onlangs op mijn werk èn thuis is overkomen: in ieder geval heb je dan nog je favorieten. 😉

De Google Toolbar is wel alleen voor jezelf, je moet ook steeds inloggen. Het vervangt de sociale functie van del.icio.us dus zeker niet. Ook kun je het niet gebruiken om zelf op nieuwe ideëen (gesuggereerde pagina’s) te komen. Maar voor je eigen werk & privé bezigheden werkt het prima. Eventueel alleen voor de meest gebruikte adressen en naast del.icio.us.

Google Toolbar

favorites 85%

bookmarks toekennen

Figuren:

Boven: Google Toolbar

Midden: Favorites (ster) en vervolgens MY places aangeklikt, Links verschijnen dan de adressen met deze ‘tag’. Dit is zeer vergelijkbaar met de “normale favorieten op één computer.

Onder: Je kunt heel makkelijk tags toekennen door een nieuw label uit een lijst te kiezen. Hier is het eerste label al gekozen (spoetnik) en kun je een keuze maken uit de overige labels als 2e woord.





Design van de toekomst

21 03 2008

DNA origami
DNA-smileys van Paul Rothemund

Wederom werd mijn oog getrokken door prachtige foto’s op het grensgebied van wetenschap, kunst en design in het NRC van afgelopen vrijdag (Warna Oosterbaan, 14-03-2008). Wat vreemde, maar prikkelende foto’s werden getoond van DNA-smileys gemaakt met behulp van nano-technologie, een bord met vleescreaties uit stamcellen op een bedje van rode kool en runderleer dat in een erlenmeyer op het lab gemaakt wordt. Ik werd een beetje op het verkeerde been gezet door de titel: boetseren met DNA en sperma. Het gaat niet om enge cloneringsexperimenten, maar om een tentoonstelling “Design and the Elastic Mind’ in het Museum of Modern Art (MoMA) in New York. Het thema is veel breder dan een futuristische kijk op de gen-technologie, het gaat erom dat een een elastieken geest (aanpassing!) en intelligent design nodig zijn om het moderne leven te kunnen volgen. Op de website is het volgende te lezen onder het kopje “People and Objects”: “Design has expanded into new fields, including the interactions between people and objects. Responsive design features objects that respond to our needs rather than awaiting our instructions. The tagging of information in our environment makes the world into a live information platform. New interfaces incorporate instinctive human traits, expanding our relationships with the objects they enable us to access. Wel weer ontzettend toevallig dat ik daar net nu tegen aanloop.moma 2

Onderwerpen liggen ook geheel buiten de biologie, zoals bijvoorbeeld cabspotting, het ritme van de wereldstad, grafisch verbeeld door een animatie van het taxiverkeer in San Francisco. Veel op het gebied van de biologie is futuristisch, maar het maken van DNA-smileys met behulp van nanotechnologie blijkt werkelijkheid (zie artikel in Nature over DNA origami (maart 2006)

Voor wie niet naar New York kan afreizen is hier de prachtige website. Goed zijn de tagging en links verbeeld met dwarsverbanden op de overzichtspagina. Elke afzonderlijke foto is getagt en via tags kun je naar andere foto’s met dezelfde tag.