[x3d-public] Web using cubic postcodes for location and web anchors fingerprints

vmarchetti at kshell.com vmarchetti at kshell.com
Thu May 25 06:40:27 PDT 2023

I think this example illustrates the shortcomings of ChatGPT, namely that it is limited in the outside knowledge that it brings to bear on a question. When I read the use case described, what it seemed to me was that the user wanted to look at a particular dated page of a metropolitan newspaper. This is already available with a few minutes web seaching and traversing a newspapers site -- the "REST API" in action, with a human client. The slowest part of the process is arranging for the payment, since newspapers generally put their archives behind a paywall, but that's a matter of economics, not technology.

In other words, I don't think ChatGPT made the judgement that this is an important use-case, retrieving recent data from the huge amount on the webpage, but is a problem that at least for this scenario, is largely solved.

Vince Marchetti

> On May 25, 2023, at 7:21 AM, Daniel Alexandre CubicPostcode <bicomplex at gmail.com> wrote:
> Hello, 
> I am the creator of cubic postcode and would like to discuss it here in the context of technologies like WebGL, X3D and also in the context of the semantic web.
> Here is my question and answer with ChatGPT about web anchors used with cubic postcodes and Unix timestamps.
> Here is a system called: "Smart webanchors" or just "web anchors". You select with a rectangle an area of a webpage that may gave text or not and that may be static or dynamic part of the page. That rectangle area is made into an image saved with bitmap format. The raw data of that bitmap rectangle and the full URL where that part of a webpage is than concatenated as raw data and hashed with SHA-512. That has is a fingerprint called a "web:anchor". It is stored in a database managed by fcc.org <http://fcc.org/> and by w3c in a special consortium to improve semantic web systems. Web anchors are 512 bits, expressed in decimal. They can be refreered with HTML tag. <WebAnchor></WebAnchor>. So webpages can embed those tags and have dynamic programming languages render them as wanted. For instance, with PHP you could use OCR to read the text in that part of the web when (and where) it is being accessed, dynamically. Here is an example. You go to a newspaper called "San francisco Daily news". You are accessing it fom cubic postcode, for example, 4444555566667777888822 at time with UNIX timestamp 1685012331. And you are interested in the first words displayed on that webpage, namely the date and time you are accessing it and where the web page is hosted. You select a rectangle on top of the page where the words San francisco, and the date of that day and first words on that section of the page are displayed. The full URL is: https://sanfranciscoDailyNewsExample.com/food/4444555566667777888822/today/25May2023/12345.php <https://sanfranciscodailynewsexample.com/food/4444555566667777888822/today/25May2023/12345.php> . When you hash with sha512 the raw data of that part of the web followed by this full URL you get the fingerprint 4dff4ea340f0a823f15d3f4f01ab62eae0e5da579ccb851f8db9dfe84c58b2b37b89903a740e1ee172da793a6e79d560e5f7f9bd058a12a280433ed6fa46510a. So if a webpage on the web has the tags <WebAnchor>4dff4ea340f0a823f15d3f4f01ab62eae0e5da579ccb851f8db9dfe84c58b2b37b89903a740e1ee172da793a6e79d560e5f7f9bd058a12a280433ed6fa46510a</WebAnchor> it can fetch the static part of the web stored at the fcc.gov <http://fcc.gov/> and W3C database of web anchors or even unsertand what part of the semantic web does that fingerprint refer to and retrieve the most recent version of that part of web, regarding accessing it at a specific date and time and accessed from a specific part of the world with a specific cubic postcode. With dynamic programming languages and javascript you are able to extract relevant information about that part of the web, namely its HTML, etc. and manipulate that data and render it up to date in real time. What do you think? 
> ChatGPT: The concept you described, known as "Smart webanchors" or "web anchors," seems to be an innovative approach to identifying and retrieving specific parts of webpages based on hashed fingerprints. By selecting a rectangular area on a webpage and generating a SHA-512 hash of the raw data within that region combined with the URL, you create a unique identifier for that specific content. Storing these web anchors in a database managed by organizations like the FCC and W3C consortium can provide a centralized repository for these fingerprints. By embedding the generated hash within HTML tags, web developers can reference and fetch the corresponding static content or understand the semantic web information associated with that fingerprint. The ability to dynamically retrieve and manipulate specific parts of web content based on the context of access, such as the date, time, and location, offers interesting possibilities. Using dynamic programming languages like PHP or JavaScript, you could extract and process relevant information from the identified content in real-time. Overall, the concept of web anchors seems to provide a mechanism for precise content retrieval and manipulation, facilitating the development of more intelligent and context-aware web applications. However, it's important to note that the actual implementation and widespread adoption of such a system would require significant standardization efforts and coordination among various stakeholders.
> Any comments for the discussion? Any feedback is welcome.
> Kind regards,
> Daniel Alexandre
> Creator of Cubic Postcode
> (feedback "at" cubicpostcode com)
> #CubicPostcode_com
> _______________________________________________
> x3d-public mailing list
> x3d-public at web3d.org
> http://web3d.org/mailman/listinfo/x3d-public_web3d.org

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://web3d.org/pipermail/x3d-public_web3d.org/attachments/20230525/94a11593/attachment-0001.html>

More information about the x3d-public mailing list