Did Twitter Help Solve Google’s Problems?
After posting about Google’s problem and my subsequent confusion, I realized that I didn’t give Twitter it’s proper due. I made mention of my tweet and the many other I saw (1200+ in the first hour), but I didn’t ask the bigger question.
Did Twitter have a hand it getting this problem fixed more quickly than it might have been?
Google Flags Whole Internet As Malware
I’m planning a trip to Italy. While I am there, I may fly back for the Mets home opener - seriously – so I tried to search for GMT to get the flight times straight.
Google warned me about every result. I was very annoyed by two things. One, I use a Mac so most of these warnings don’t apply (I KNOW some do), and 2) the only way to continue to the site is to copy and past the URL (or type it) and go.
Then I noticed that even the Wikipedia entry was flagged. What?
Then I decided to search to see if my websites were flagged – Yep! Even my twitter page, hmmm? So, I searched for Twitter – Yep.
What about YouTube, a Google property. Searched for youtube and sure enough the first result was www.youtube.com which was flagged.
So, I sent a tweet to ask for help. Seems that about 300 people sent tweets just minutes before.
Good news – the problem seems to be fixed.
Google Flags Whole Internet As Malware .
I haven’t eaten since later this afternoon
I love the idea of time travel and am fascinated by the paradoxes it may or may not cause. Primer is a movie that addresses these paradoxes head-on. It’s a must see movie for anyone who likes time travel. Hang on, it’s a wild ride.
Since the ladies are away for a few days, Billy and I decided to have some good sci-fi stuff around to watch. He already had Primer on hand and four eps of
The X Files
After watching it on Saturday, we found The 10-Step Method to Watch Primer and Really Understand It.
1. Watch the movie: if you get really confused – great, that’s part of it.
I’ll say we were confused! That’s what sent us looking online for explanations. Now we knew what to do next.
2. Watch the movie again: still confused? Good. You did step 2 so you can better understand step 3.
3. Look at this timeline chart: http://neuwanstein.fw.hu/primer_timeline.html
Read over it twice.
The comments point out four other timelines and we reviewed three of those (fourth no longer available) too. Billy prefers a more narrative understanding of the movie and preferred this one. I like the dates and times so I preferred this one, which is low on speculation and very succinct. The third is an interesting take loaded with speculation – fine with me. 
Can’t move on to step four until we get a DVD copy again. We checked a local store and didn’t find it so I’ve ordered my copy. Amazon’s DFC is just a few miles from here so it’s actually posssilbe I will get it tomorrow (even though it’s Sunday), but we can wait a few days if we must.
I’m sure I will have more to say about this one another time.
Go, watch it! There is a “ten step” program for you if you need it.
Daily Links
Twitter Search Ranked By ReTweets
The existence of any ranking system based on retweeting, officially sponsored by Twitter or not, will have a negative impact on the Twitter experience. It will change the way people tweet. IMO, that is a bad thing.
Yesterday I posted about filtering Twitter search results and I left off with a quote about Google’s PageRank which I think is apropos to the discussion of ranking tweets based on the number of retweets (or twitterers based on the number of tweets being retweeted – or sheer number of retweets from a twtterer)
PageRank stopped working really well when people began to understand how PageRank worked. The act of Google trying to “understand” the web caused the web itself to change. -Jeremy Zawodny
The importance of this cannot be understated. Google, once a company which wanted to provide relevant search results by understanding the relationships between webpages, is now a company that dictates the decisions of millions of webmasters. They don’t do this by decree, it is simply a function of people’s desires to be found on the web.
Google’s idea was a good one – use the structure and nature of html pages to rank them in terms of importance (“authority”) and then use that as one criterion for determining the relevancy of a page. This good idea led to at least two inherent problems.
First, it lent authority to webmasters who deserved no such thing. Years ago, I had several web pages with decent pagerank. As a result, I could post a page about almost any subject and my page would rank higher (perhaps) than a real expert on the subject. As I noted yesterday, others have pointed out this problem. Of course, the web being what it is, my page could be usurped by the proper authority on the subject – IF the web behaved as it used to. This brings us to the next problem.
The web now behaves differently because of Google. I am not complaining about this. I am not unhappy about Google’s existence – quite the opposite. I am simply making an observation that many others have made before me. Webmasters now have some understanding of PageRank and now decide how to construct websites to optimize it.
A retweet ranking system will cause the same thing to happen with Twitter. It will undoubtedly cause people to change the way they tweet. Many users will:
-
think twice before tweeting something that is not “retweetable”
There go all the fun updates about feeding the kids, eating breakfast, etc. I know many of you don’t like those, but I’d argue that most users enjoy this Twitter “nonsense.” -
compose more retweetable tweets
Now more tweets will be like those annoyingly profound “updates” on Facebook – yuk! -
ask for retweets
I propose that any ranking system automatically take this into account and not credit for any tweet which contains the phrase “please retweet” or similar. -
retweet less
Many will hesitate to give away authority to others, or at least be more careful about who they give it to. This is very similar to PageRank leak. -
seek more followers
While number of retweets is certainly better than a straight ranking based on number of followers, it is still very much a function of the number of followers. Ideally, those with more followers need to be retweeted less but I assume that doesn’t hold up in practice. More followers will probably lead to more retweets.
Obviously, I am assuming a simple ranking system which rewards users whose tweets have been retweeted. I understand that a system can be put in place that ranks the tweets and not the twitterer. I’d argue that all points above still hold. I also understand that a more complex formula can be devised, but the point is still that people will tweet differently. Imagine a system that rewards one for retweeting other’s tweets. This will lead to a Digg-like echo chamber where perhaps very few tweet original content.
As I said in yesterday’s post, I am in favor of filtering my Twitter search results. I’m just not a big fan of ranking by retweets. More on that later.
Twitter Search By Authority
Loic Le Meur proposes that we need a “search by authority” feature for Twitter and then recommends that number of followers be used as an indication of that authority. Michael Arrington agrees.
Both took some heat for this proposal. Most of it seemed to revolve around the idea that authority equals number of followers, but I think the discussion is more interesting than that one issue.
I can think of at least four questions this raises.
Do we need to filter search results at all?
I think the answer is yes – sometimes.
Gary Arndt comments
Microblogging doesn’t lend itself to search. Period.
I agree, but I think he is overstating a bit. A Twitter search is not the place to do in depth research on a topic, but the fact that the results are “all over the map” doesn’t mean that one can’t find useful results. The timeliness of the results alone can be very useful. There is simply no doubt about this.
Several commenters pointed out that we should all just enjoy Twitter the way it is. I understand the argument here but there is nothing wrong with discussing new ideas and improvements to existing technology. I enjoy Twitter the way it is and plan to use it the way it is, but I don’t mind that others would like to make it more useful.
Most of the time, I want my Twitter search to return more results than I can possibly process. There may be times, however, that I’d rather filter out some results. Having a mechanism to remove those results voluntarily is always a good thing. I use the minus operator all the time to narrow down my searches. If I could filter based on other criteria voluntarily, that would be great too.
Should we filter based on authority?
I hesitate to answer this honestly because I think many of you will stop reading here,
but I think the answer is yes.
When I read Loic’s post the other day I was as irritated with his tone as many of you, but I do think his point is a valid one. What he was recommending, according to his clarifying comments, is an option to filter based on authority. While I might not use this feature as much as some, I think it’s a good idea.
I tend to agree with Steven Walling who comments
for many, many people, the point of Twitter search is to find things said about a topic of interest regardless of how influential the person speaking is.
and Joe Duck, who adds
In fact I think it would be interesting to reverse the algo you suggest – I’d rather hear from some Grandmas in Peoria about their iPhone experiences than from Jason Calacanis about [groan] the wonders of Mahalo.
I could not agree more with these two comments. Our family had so much fun on the night of Hillary Clinton’s concession speech.* Terri and I both punched slightly different search criteria into Twitter Search and we read some of the more interesting (and family friendly) comments aloud. It was a good time had by all. None of us cared one bit if the people spouting off had any kind of authority at all. That wasn’t the point.
I may not be interested in search by authority, but I understand why some people want “authority” opinions on a topic from time to time. Some of these might be egomaniacal, but that should not keep us from using an “authority search” for legitimate purposes, should it?
This brings us to the most difficult question of all – though perhaps the least interesting to me personally.
How does one measure authority?
I think the verdict is clear from the weekend discussion – the number of followers is not a good measure of authority. Many people weighed in on this one but two comments sum it up for me:
And if authority is what we’re after I doubt if number of followers equates to authority. Too many really smart people have very few followers. -Dave Winer
Popularity does not equate to accuracy. If it did and Twitter existed 2000 years ago; the earth would still be flat. -Josh Jonte
As far as I’m concerned, that settles it. I’ve used Twitter since 2006 and my experience tells me that the number of followers is not a valid measure of authority.
Nitin Borwankar and Taylor Barr also touch on the problems with such an approach and hint at the possible solutions. I will have more to say about those later.
So, is there some other measure one can use to judge the “authority” of a tweet or twitter user? Scoble and others recommend we use some other metric – retweets, followers/following, someone even proposed
~= #followers/(#faved + #RTs)
My interest in finding the perfect formula for this is negligible but I do think that we’ve discovered one other measure that won’t work – retweets.
Once again, I think two comments sum it up best:
A sycophantic echo chamber is as stupid of a metric as # of followers. -T. Luoma
All these popularity contest-like sites and lists full of people echoing each other all the time just seems to take some of the original spirit away from Twitter. imho. -kosso
Michael Arrington points out that “This is exactly what Technorati does with blog search” and Robert Scoble points out that this is just what Google does. This, by itself, is reason to question it in my opinion.
Scoble goes on to point out that according to Google he is the number one result for “recession 2008” “How did that happen?” he asks.
Jeremy Zawodny points out a similar issue in 2003 and correctly (IMO) diagnoses the problem a few months later:
PageRank stopped working really well when people began to understand how PageRank worked. The act of Google trying to “understand” the web caused the web itself to change. -Jeremy Zawodny
I will pick up my thoughts on this tomorrow, but I do want to bring up the last question.
What other filter criteria might be useful?
It seems clear that at least some users would like to filter search results from time to time. What criteria would be useful to you?
* Terri and my daughter were huge supporters of Hillary Clinton. While my son and I supported Obama from the beginning, we would have wholeheartedly supported Hillary if she’d earned the nomination.
Kobe vs. Wilt
On January 22, Kobe Bryant scored 81 points in a single NBA game. This is an amazing feat which one can’t help compare to Wilt’s 100 point game. A Sports Illustrated article which reminds us of the difference in stature between the two players even prompted some to hail Kobe’s performance as superior given he is six inches shorter than Wilt. While I think this is a ridiculuos idea, I couldn’t help but put together a spreadsheet to determine what it takes to “break the record.” I factored in three point shots while I was at it. Of course, I couldn’t resist an online version as well.:)On a related note, I read another interesting article about a man who gave up tickets to see Wilt’s 100 and Kobe’s 81!
RSS Readers
In a previous post I rambled on about my own personal search for the right way to manage the RSS feeds I want to read. Regular readers will remember a similar posts about link managers. Read more
del.icio.us, furl, my linkblog etc.
I’ve been fooling around with how to post links to my linkblog and del.icio.us for a while now. I posted about this three times in July (14th, 23rd, 31st) while I was playing with the best setup for me.Since then, I have toyed with a few different tools for managing this. In September, I developed a solution that I liked. I wrote a toolbar button which, when clicked, would take the info from the page I was on and post to my linkblog using a combination of builtin WordPress code and my own code from 7/14 and then post to deli.icio.us and redirect me to deli.icio.us “done” message. Aside from the occassional unescaped character, it worked beautifully for a while, and I liked it.Recently, I noticed that the links were not posting to my blog (I know some of you don’t understand why I even want this – that’s okay). After messing with it for a few minutes I decided to just do things another way. I wrote another toolbar button that would post to my linkblog using builtin WordPress functionality and popup (in a tab, of course) the del.icio.us post form with the tags populated. I copy the tags, submit to del.icio.us, paste my tags into WordPress, and submit. Quick and easy enough for me.If you use FireFox and WordPress, you can drag this link — Post Link to your toolbar.Now, edit the code as follows – replace DELICIOUSUSERNAME with your del.ici.us username and replace WORDPRESSURL with your wordpress blog URL.
E-Mail turns 34
Official Google Blog: Guess what just turned 34?I remember when I first disovered e-mail. It was 1979. My parents purchased my brother and I a TRS-80 color computer, and a 300 baud modem. Of course, we used the modem to dial our friends directly to ‘chat’ and to connect to a couple of very cool BBSs. The computer did come with an introductory membership to compuserve which included e-mail. The connection was $12.95 per hour, so we never used it after the free trial ran out, but I remember thinking how cool the idea of e-mail was. Remember those e-mail addresses 7,12345,678@compuserve.com.Almost seven years would go by before I actually sent an e-mail to someone I knew. Of course by the early nineties I was using it all the time. And now – I get 1000’s of e-mails a day. I still love it, though!BTW, I still keep in touch with many of those friends and the guys who ran the BBSs.
Information Overload
I’ve finally decided how to manage all the content I’d like to consume – I think
Until yesterday, I was tracking 341 feeds in my bloglines account. I’ve had a base set of about 310 for quite a while and I add new feeds all the time to a test folder. I was having difficulty feeling in control of all that info for a couple of reasons.
One – an immense volume of information. Many of the feeds I read generate many posts. I cannot wade through all the headlines every day and therefore I become swamped with even more info on the next day. I would frequently fire up bloglines and find the number of entries well into the thousands.
Two – am I really in control? I was beginning to feel like I was being pushed around by the UI at bloglines. I like the look and feel generally, but it can be a little sluggish, especially when I open a folder with 1000+ entries. I wouldn’t call it slow, but it could be a little slicker.
Looking for a solution
More than six month ago I began to look for a solution to my information overload problem. I played with Thunderbird and live bookmarks in FireFox; but I decided to stick with a web-based solution of some kind. I love the fact that I can reach my feeds from anywhere I can find a browser. Obviously bloglines was in the running and I decided to try Rojo a lot too. The post-centric flow was a nice feature for me and the website is sleeker but it wasn’t always very fast. In the end, I was just afraid to switch and too lazy to try and work both for too long.
Along comes Google’s reader
Very similar to Rojo with the added benefit of *nix style keystrokes. This was a big plus for me. It was one of the things I liked best about Gmail when I first used it. It used keyboard commands that are very similar to mutt, vim and some of my other favorite apps. Two very annoying features forced me to give up right away however. I have read that there are reasons for some of the problems with Google Reader.
Bloglines decides to use hotkeys
Some think bloglines jumped to this conclusion based on Google’s entry into this space. Good move on bloglines part. I am certainly not the only user who prefers to control my applications with the keyboard. Regardless of the reason, this made me feel much better about my decision to use bloglines.
Wait a minute.
I still have not solved my problem with overload. With all the excitement (at least one user is not as enthusiastic as I) over bloglines announcement of hotkeys, I’ve forgotten to deal with the real problem. While it’s on my mind, I decide to just settle this issue once and for all. Then it dawns on me. Use one for news and one for blogs and the like.
You see, I’ve actually been looking for a way to get my eyes on even more news information. As part of this goal, I had considered using another tool for this. So this all seems to have worked out for the better. Bloglines uses folders and a folder and feed centric interface which, I think, works best for blog reading. Rojo uses tags and has a post centric interface, which makes sense, to me, for news. This also eliminates a huge problem I faced with my old setup. If I happen to fall behind on the news, I can catch-up. I was always very reluctant to use mark all read with bloglines because I didn’t want to miss anything from the blogs. Usually, if you miss news, you’ve missed it. No need to read old news, right?
I exported my subscriptions from bloglines, edited the OPML to include only news sites, and imported them to my Rojo account. Then I unsubcribed from those feeds in bloglines and deleted the folder. Now I can manage to monitor even more feeds and feel better about the whole thing.
Google Starts Up Philanthropy Campaign
McGurn said that at least Google disclosed to potential shareholders up front that it intended to fund philanthropy through the corporation. But, he added, there is a danger that the “lines get blurred between how founders view corporate assets and personal assets.”
Google Starts Up Philanthropy CampaignI’m confused by this whole thing.Many companies give like this. Many stockholders support it. Many don’t.I am not sure Google’s founders are ‘blurring’ any line. I think they know what they’re doing. Stockholders have every right to speak out. If enough stockholders disagree, fine. Until then we should all wish other companies would aspire to do good.
More About Links
I haven’t posted too many links the last couple days. I have a bunch of guitar links to post – like I said before, my links come in waves , but I’m not sure I want to post them all. Some of them generate popups. I have nothing against the responsible use of popups, mind you. Some sites rely on them for revenue. I used them myself on the first two web projects I created (here and here – Both are under different management now.), but I don’t like the way some publishers use them.Anyway, I did a search Read more
Link Managers
I’ve been using Furl and del.icio.us for some time now. I use them in combination much like recommended by Amy Gahran and Saurier Duval. I used de.lirio.us instead of del.icio.us at first, but have moved to del.icio.us. It’s been more reliable and I like the to be able to grab my links via the api. I post them to my link blog using a daily cron job.
About a month ago, I ran into this post by Michael Wexler 4/28/2004 (who hates tags) comparing some other services. I decided to search for more info and see if I had missed a better site. I found a good list by Clay Shirky, and a great set of reviews by Kossatsch (version 2, version 3, and pdf chart). I’ve played with most of these sites over the last month and I’ve decided to stick with the system I’ve been using.
I’ve had fun though, and did some interesting reading about tags, tag literacy, folksonomies and controlled vocabularies.
Google Maps
Tell me this is not cool!The night it was announced my son and I began brainstorming what we should try. We came up with about two dozen ideas and decided I should try about 8 of them.Here are two maps I threw together as a learning experience. They are not finished yet, but it shouldn’t take much moreUPDATE: I forgot to post the links to the maps. This one actually works. It may not work well, but you can actually do searches. I have to click on one of the panel items to make the map show, but I am looking into that. This one is a mock up of the next one I’d like to finish. It doesn’t work yet. I plan to pull all the data and store in on my server.
My Link Blog
A while back I toyed with de.lirio.us for my link blog. I pulled the feed and transformed it into html. I used the html from my blog so it would have the same look. I was becoming less satisfied with de.lirio.us and started itching to use del.icio.us instead. I thought I could pull the feed and post each item to a real wordpress blog by dropping the data into the database. My solution is not pretty, but it works. I run a cron job daily to do the ‘pull and post.’I like it. I’ve poste the code in case you’re interested. Read more
The Only Winning Move is Not to Play
While doing laundry, I noticed that my shorts and tees are wearing out badly. Later, while reading some of my favorite feeds, I found this. Thanks for the tips. I will absolutely pick up the SQL shirt and M.A.D. I’m going to shop for more too. Keep an eye out for my wishlist in the sidebar.
Code Monkey
I spent several hours experimenting with Amazon Web Services yesterday. We do most of our shopping online so Black Friday
is not a big day for us.I have been working on a couple of projects with AWS. If you have any interest in such things, you can read more.
Wow, that was easy!
I have just decided to switch from Blosxom to pyBlosxom. Mostly because I use Python in my classes. I like Perl a lot and I really like Blosxom, but I wanted to try this. Frankly, I did not think it would be this easy. In my experience, any Perl script I grabbed from the web has always worked more ‘forgivingly’ than any Python script. Now I just have to find (or write) the right plugins.I will not be moving old posts here.
Using BookMarklets
While doing some research for my next project, I ran into some JavaScript code that intrigued me. I don’t use JavaScript, because a long time ago I was brainwashed into thinking that JS was never to be used for anything. I have always thought it was ‘cool’ but thought that it was bad design to use it on the web. I remembered vaquely (from Simon Willison’s Blog as it turns out), reading about something called bookmarklets. After a quick Google for bookmarklets, I decided — “this is what JS is best used for!” I added several bookmarklets to my links and now I can use them to navigate websites, analyze website designs, and much more. The Google link above will take you to most of these but my favorites are:
-
Subscribe
-
What I'm Reading
-
What I'm Watching
-
About This Site
This website is simply a place for me to express my opinions. I post my thoughts on faith, family, freedom, and fun. -
About Me
Subscribe via RSS
