In less than 24 hours since I have published the proposal for “about microformat” to tag media on the Internet a number things become obvious. Firstly, there is a very poor understanding of the Internet as media platform in general amongst the creative community. Secondly, there is even less comprehension when it comes to the Internet indexing by search companies such as Google and Microsoft and subsequently how the search results are or can be formulated by these companies. There is also a third issue which confuses many people, namely proprietary tags associated with specific media, such as EXIF generated by modern digital cameras.
I will try to expand on these subjects in hope that greater level of understanding amongst the creative community will lead to a quicker adaptation of the about microformat proposed by me, or a development of an alternative tagging methodology.
Internet and your media.
When you publish any media on the Internet such as photographs, music, paintings, drawings or even Flash games your browser has built in support to recognize them and present them to you, when you view the web page they are attached to. Your browser however does not know what they contain. It uses HTML tags inserted by a web developer, web designer or even automatic programs which helped you to create your web page to render the content. In case of photographs, your photo of your ginger cat with a file name of dsc00095.jpg will end up coded in your page as
<img src=”dsc00095.jpg” />
If you have bit more of Internet and HTML knowledge, you will rename your photograph to let us say “ginger-cat.jpg” you may end up with something like this:
<img src=” ginger-cat.jpg” alt=”photograph of my ginger cat named Ally” />
If you are dedicated, you can turn this into:
<img src=” ginger-cat.jpg” alt=”photograph of my ginger cat named Ally” longdesc=”Ally, my ginger cat I photographed during Christmas Holidays.” />
None of the text you see in above tags would show on the Internet so when you publish this photograph to share with friends you would normally write a paragraph of text explaining what the photograph contains. For example, you may repeat the content of longdesc attribute: “Ally, my ginger cat I photographed during Christmas Holidays”. Your friends and family who know where your web page is will be able to enjoy the photograph and can read the story associated with it.
Your media and search engines
So, what happens if you publish 1000 of photos like that and you need to find a specific one or you would like to learn how others photograph their cats? You turn to search engines such as Google or Bing for help. You type cat in the search field and press the “find” button. If the photographs were embedded in pages using the first method you will never find the photograph as dsc0095.jpg no matter how hard you try to read it does not spell “cat”. You have even less chance to find a ginger cat from that filename. Photographs coded in the second way and even more so in the third way have reasonable to good chances of being found … eventually. However if someone searches for “feline, Burmese, photograph” they will never find your photograph, even though Ally indeed is a Burmese cat. You could go back to your page and add more information about the photograph, but at some point, the search engine will have so many keywords from your text it will no longer see which ones are relevant to your image. If you are not a good writer your photograph may eventually be found under “Christmas Holidays” and not under “cat”. We humans are very good in recognizing and categorizing the information we see and read, but this is not case for computer programs.
All this happens because HTML, the language all Internet pages are created with, does not contain a semantic way of tagging your media.
Search engines and media specific metadata such as EXIF in digital photography
Some of you, familiar with specific media formats, may say “OK, but my camera inserts EXIF information into the file and this covers 99% of what is required for search engines to discover my photographs.“
EXIF is not part of HTML and is not indexed by search engines. Search engines, which index your pages, read only the text on these pages. Your images may be indexed very slowly, many months later. They will be tagged by humans to reflect what content they contain.
EXIF does not contain semantic metadata. If you take a photograph of a cat, your camera does not insert the word “cat” in there, nor can it discover it is ginger, Burmese and its name is Ally. Nor does it know that the photographer’s name is “John Doe” who decided to place the image into the public domain. Additionally many media formats do not contain EXIF at all. A simple example here would be an older film photograph, scanned into a jpeg file.
Did you know that for music, video, and Java and Flash games creators the story is even harder as their works are referred to on the Internet simply as “object”.
This is why I am proposing the “about microformat” an easy to understand and implement method of describing any media you publish on the Internet.