Image Search

Mount Kazbek church, Republic of Georgia
Mount Kazbek church, Republic of Georgia

Google has a cool tool, Google Images, which can search images. Provide it text, and the images returned will have related metadata or page information to your search. Now, for the really cool part, you can search based on another image.

Click the camera icon in the search bar and another box appear. Enter a URL or click the link to upload one. It uses the image provided as the search and returns similar ones.

Some uses I have for it..

  1. Who is using your images. It is easy for someone to download any photo posted on a web site. Then they can upload it elsewhere under another attribution. Searching for your images can help locate someone who is re-using your work.
  2. Correctly attribute images. I see a photo without identifying information and desire to find the source.
    1. Painting. Maybe a painting and I want to see more of the artist’s work.
    2. Photograph. Ditto. A concrete example is I saw a background of a web page for a State of Georgia (USA) web site with a Russian-style church with mountains in the background that looked nothing like those in this state. Searching on that image turned up a Blogger page with the same photo identifying it as in the Republic of Georgia.
    3. Identification of plants, animals, etc.
    4. Locate higher resolution version.
  3. Finding similar work. Once you click into “Visually similar” photos, you have all kinds of neat controls like size, color, type, and time. Maybe a logo looks derivative, but I am not familiar enough to know. Image search can locate very similar logos and point to the original.
  4. Scams. A friend was renting an apartment in Amsterdam and wanted to know if the place was legitimate. Using the photos from the email, I was able to find multiple other listings that all used the same photos.
  5. Identify Fake Profiles. Scammers are lazy and take photos from elsewhere on the Internet. This can find the original.
  6. Debunk Social Media. People share doctored or misattributed photos on social media sites all the time. This can find the snopes or other anti-urbanlegend site’s page on the photo.

I am sure there are more.

Anyway, I use this at least once a week.

Curious Traffic Spike

I glanced at my Google Analytics stats for this site and noticed a huge traffic spike. Somehow my TED Talk: We Are All Cyborgs post landed Bing’s number two spot and Google’s number three spot for “ted talk we are all cyborgs” a couple days ago. Normal for a Tuesday is something like 650 visits. That Tuesday I got 2,578. It kind of reminds me of the Made Stumbleupon.com? post.

The actual We Are All Cyborgs talk was the number one spot for both search engines. Why would anyone come to my site for the same video?

(Glad I turned back on WP-Cache again.)

Fix WP Numeric Broken Permalinks

In the early days of my using WordPress, I set the permalinks setting (the URL format style) to Numeric. They looked like http://ezrasf.com/wplog/archives/3. On 2008-SEP-27 I changed the permalinks setting to the Day and Name. According to my broken links post each time WP autosaved a draft it incremented the number so the names were no longer sequential. The gaps annoyed me. The new setting hid those gaps. (No, I do not have OCD.) However, it meant

  1. all those links in posts to old permalinks no longer worked and
  2. anyone incoming from search engines hit permalinks who no longer worked.

The search engine problem worked itself out without any effort on my part. They recognized the 404 HTTP error code, dropped the bad page from the index on the next crawl. They also picked up the new posts.

I occasionally spent some time working on fixing broken links. However, the process of determining where the link should go took so much effort I rarely fixed more than a few links at a time. So I did not make the progress I would have liked.

Then I discovered the Broken Link Checker plugin for WordPress last weekend. It has been sending me notices about all my broken links. In desiring to shut it up, I had to spend time trying to fix those Numeric permalinks again. I noticed the format of a link in “Get Shortlink” buttons when I edit a post is the same as the Default permalink which look like http://ezrasf.com/wplog/?p=3. It seemed logical I could just replace “archives/” with “?p=” and fix the internal links. Sure enough, it worked. So I’ve cleared up the remaining internal broken links much more easily than I ever expected. It could only be easier if the broken link checker automatically did it.

The WordPress Codex says,

Find a post’s ID number and type the following (with your information) in your browser and you should be redirected to your post:

http://yourdomain.example.com/post/(the ID #)

Well, no matter what id number I use here, they go to the same post on 9/12. Weird. This would have been an even easier fix as I could double clicking on archives does not get the slash. Maybe it means I need something in the .htaccess to make it work correctly?

Meh. I am glad to have an easy solution. Annoyed it seems undocumented. Hope this helps someone else who has the same problem.

Expression Costs

(This started out as a blog comment for Sania’s post Facebook Killed Your Blog. I’m posting it here first.)

We share blogs with the whole world. So our blogs get lost in the noise, bolstering the need for a whole industry optimizing getting found in search engines. Its a concerted effort just get noticed. That’s because blog readers have to seek out blogs to follow, subscribe to the feed, and follow. Finding the best blogs to read is sometimes difficult and more from word of mouth than anything search engines provide.

Blogs also tend to have a lot of information to digest. Social networks have just a line or two with maybe a link to more information. Blog readers typically are designed around the idea of collecting all the posts and letting the user pick which to read. Social networks typically are designed around the idea of just showing recent posts and letting the users choose how far back in time to read.

As technologies lower the costs to express ideas (aka get easier), blogs will get left behind as they have become upside down in value. The costs of writings, reading, subscribing, and commenting on blogs are more expensive compared to micro-blogging or status updates.

Why blog when hanging out on social networks are so much easier? Blogs can only survive as long as they have information worthy.

Why blog when readers are no longer reading? Posting blog entries on social networks does help keep traffic levels somewhat by getting exposure.

As bloggers providing valuable expression leave blogging, the value of blogs decrease. People will still blog. It just won’t be the popular thing to do.

Who Are You?

I’m so vain…. I probably think this post is about me….

Probably only people who do vanity searches notice this, but there are spiders pulling names off web sites. They link the names to companies, blogs, and other web content. Supposedly, these sites allow online reputation control. Rather than you claiming your identity as others in this market, they list you in their database with the hopes you claim it.

See, you probably have accounts on several web sites. The idea is to both aggregate the accounts and prove ownership. If your name is John Smith, then you probably are getting confused with other John Smiths. You’ll provide where you work, contact info, which sites belong to you. The site will provide a feed showing your activity in each of these.

My name is pretty unique. If you saw my full name on a site, then would you doubt that its me. Okay, let’s forget the guy who masqueraded with my name a few years ago. Lots of people say I have the best names. He took it too far. By contrast, there are others with my first name who pop up higher in Google. So, you’d need the whole thing. I notice people arrive at this site by putting that name in search engines, so I am pretty sure it works. Naturally, all the sites where I wish to stay under the radar don’t have my name on them.
🙂

See… I knew I’d make this post about me.

Am I the only one who remembers fascination with the Deep Web (aka Invisible Web)? The idea of these online reputation services, I think, is to bring positive content up in rankings up to the more shallow areas. Trick is, the users need to be aware of what is and is not positive. Linking your name to your Facebook (used to be Deep Web but less and less of late) profile and giving the world access to pictures of you passed out drunk probably isn’t positive online reputation control.

No More Public Anonymous Photos

Early in the history of the WWW, people thought a document which did not have an inbound link, a link from some external location to the file, was private. Search engines looked for content on web sites in locations for which there was not a link already pointing. To truly respect the privacy of those sites, these search engines would have only index content from sites which asked to be index AND only indexed content which that web site pointed a link.

Just because SiteB links to a location on SiteA doesn’t mean SiteA wished it to found in a search, an example is deep linking. Note the court cases (1, 2) in which judges rule in favor of the SiteAs who go to court about SiteBs who make such links. Note I made deep links in order to demonstrate the deep links. 🙂

On Privacy and Polar Rose – Polar Rose Blog : On Privacy and Polar Rose

It should come as little surprise that we believe that Polar Rose adds tremendous value to the photo web. We think we’re as harmful to the photo web, as Altavista, Yahoo!, and Google have been to the text web. By sorting the text web, these search engines exposed the wonderful resource of public documents that web had already become. The side-effect was that information which was not meant for public consumption, but which was kept private by obscurity, was suddenly exposed and searchable.

By Polar Rose’s logic, because people acclimated in general to losing textual anonyminity, they will do the same for facial recognition. Just what does the lack of a label mean? The photographer may be protecting the identity of someone, a minor for example. The uploader may be lazy. The uploader may not know. The uploaded may not own the copyright to the photo.

On the whole, I think facial recognition is a good thing. The cases in which there are likely going to be privacy concerns are going to be more likely uncommon than common. Those who it affects are going to be most upset all the same.

If anything, then I think we have a tendency to underestimate how badly such things are going to violate the privacy of our lives. For example, look how often search engines exposed Social Security Numbers on web sites.