22 March 2005

How to convert Haloscan comment codes to titles, even when prohibited

I’ve noticed something on Feedburner. The titles on popular content are sometimes meaningless. Not because the authors choose a bad title. But because the original platform issues codes, not words. What use is that?

For example, Haloscan reports on an XML feed a comment-number as the feed title. Then, Feedburner reports this number this as the title. A list of codes in Technorati isn’t interesting. Why not use plain language?


Workarounds


You can modify your-name to include many words. When Feedburner posts the link, the message will have some plain langauge. But the code is still at the beginning. Just like everyone elses. That does not stand out.

Also, you can put HTML code in the name-box, but the title-code is still there. Oh, and the HTML may not work. That's no use.

How about this. Hide text in a Google search, then copy the URL and convert it to a tiny code. But this is complicated. Even if it works, it's not much use on new topics with unpublished terminology.


Constraints


The goal of this is to have meaningful titles in both platforms.

For the sake of argument, let's assume Haloscan does not change and still publishes the code. Let's not rely on a search engine for this workaround.

This means the solution means embedding a lot of text-information in a small Haloscan box; and then have that information magically appear in a second platform like Feedburner.


What's an easier hack?


Users should be able to use the smallest space to have the longest title.
How can I convert a Haloscan code into plan language and more meaningful title in either Feedburner or Technorati?
I came up with three approaches, none of which I have seen, but build on the above concepts.


Quick overview of the approaches


XML Burst: Converts a long-code-title into alpha; saves this code; assigns it to a second platform.

XML Burst Mouseover: Takes a tiny code, assigns alpha characters; user when they place their mouse over the title can see the embedded title and other details.

XML Title Extractor: Takes title out of the content, assigns a tag to it, then associates the comment-feed-code with this title code.



XML Burst

Converting code to alpha and compressing text into small spaces


Consider the URLs today. They used to be numbers. Then they added letters. Now, users can crunch these alpha-URLs into snipped or tiny URLs.

We can put many URL-characters in a single-search-URL using a snipped URL. So too should we be able to compress a lot of plain text into a small space.

It would be nice if we had a special code that we could use to embed the text, and then have the second platform know to convert this code into a larger sentence, paragraph, or title.

When we do a normal search, we type text into a box, the search gets completed, and then we see a URL list.

XML Burst does the opposite. It takes the tiny URL, converts those to words in a box, then extracts and applies those words in a second platform.


Modules to execute this approach


1: Input text

2: Save text

3: Converts text into a single code

4: Index saved codes

5: Integrates with an external request, deliver the code to that external platform

6: Compares the output-text with the input-text;

7: Verify input code matches the output code;

8: Verify input text matches the text that is finally reported in the output platform.



XML Burst Mouseover

Reporting detailed text with mouseover


This approach does the same as above, but rather than reporting the code in the final platform, this approach would simply report the Haloscan code. Users would use the mouse-over to see the detailed information.

This approach is less optimal for a number of reasons. First, it doesn’t readily identify the code as distinct; users have to use the mouseover; and it doesn’t really address the problem of being unable to differentiate the titles.

Second, this approach is problematic for external platforms that may index the Haloscan comment-feeds. The original problem was that a platform like Technorati that is indexing Halocan comment-codes doesn’t have the tools to differentiate between title 1 and title 2; the only difference is the users-search request, that may or may not hit content.

Thus, the first approach is better than this XML Burst Mouseover. However, this approach should not be disregarded. This approach could be integrated with the first, in providing even more detailed comments underneath the mouse-over. Again, the first option differentiates the Feedburner-titles; yet a mouseover could either provide text or a screen-shot of the actual webpage/blog-spot.


XML Executable Code

Compressing executable code in a small space


Rather than let ourselves to be confined to a given space, use the small space as a constraint that will create a tool that will allow us to have a large number of commands embedded inside a tiny URL.

Let’s presume the tiny-URl concept that applies to URLs could also be applied to text. The next step would be to apply this principle to code.

For example, the embedded code could be a conversion from HTML to Java; or a tiny URL that links to an external platform that contains the commands.

This also is similar to the idea of wireless communication of instructions about code across a platform, then using a conductor to organize these existing elements into new patterns, thereby eliminating the problem of transferring code across platforms.

The idea of a tiny URL is to fit a lot of URL-characters into a small space. So too are there times when we might want to have a lot of code to fit into a small space.

Rather than require service platforms like Haloscan to provide additional boxes, one workaround is to use a Tiny URL that fits into the name-spot, then use the tiny-code=URL to embed a larger command that will include Java code for a more complicated interaction with Feedburner.


XML Title Extractor

Associate a comment code with a name


A third approach to the Haloscan-Feedburner title-integration issue is to use a title tag. Again, this is not duplicating what already exists.

Currently users can type in a title-command after href to create an embedded title-link that appears with mouse-over.

This goes one step further. Rather than simply relying on a rel-command for tags, this approach could rely on a new code-standard, namely: Create a rel-like-command for a title-to-code conversion.

At first glance, this approach would be more code intensive than the above two approaches, but it also has the potential to have far wider applications to new platforms.

The issue remains whether the future XML platforms would benefit from such an approach. I believe there is the same kind of potential for this approach as there are for tags.

This is another way of saying that we could develop a standard that could automatically take text out of document 1, assign it a title-command, and then use that command as the default title in platform 2.

This would solve the Haloscsan-Feedburner integration issue: It would make the Feedburner title more meaningful; it would assign to the Haloscan platform and title a more meaningful embedded code; and users could differentiate the title, even though it was simply a bunch of numbers.

Also, external platforms could rely on this title-command to differentiate between Haloscan-comments and make the content in the feeds more accessible with searches.

To make this approach work, there would have to be some discussion on how to convert the Haloscan-title into something related to content; then agree on that code that would work in all platforms.


Summary


You've got three concepts to make more meaningful titles. These could be applied to Haloscan. And they could make some easier to read titles in Feedburner.

  • Option 1, XML Burst, I outlined the modules to make this work.

  • Option 2 appears to be less desirable, but the mouseover feature is interesting.

  • Option 3, XML Title Extractor is designed to create a standard that would extract content, assign it to the title, then use that title as a way of indexing in a second platform.

    Option 1 modules may be of interest. In the long run, I see more potential in applying Option 3 to other areas.

    To address the immediate issue of making Haloscan-titles-in-Feedburner more meaningful, I leave it up to the creative xml-developer minds.