Incoming links: PageRank
PageRank was apparently inspired by it originally, given that.
How to develop Ciro Santilli's website before the OurBigBook migration Updated 2024-12-15 +Created 1970-01-01
The website moved from AsciiDoctor to OurBigBook Markup in 2020, making this section mostly useless. But hey, history!
Ciro's website is powered by GitHub Pages and Jekyll Asciidoc.
The source code is located at: github.com/cirosantilli/cirosantilli.github.io
Build locally, watch for changes and rebuild automatically, and start a local server with:
git clone --recursive https://github.com/cirosantilli/cirosantilli.github.io
cd cirosantilli.github.io
bundle install
npm install
./runSource:
./run.The website will be visible at: localhost:4000.
Tested on the latest Ubuntu.
Publish changes to GitHub Pages:
git add -u
git commit -m 'make yourself look sillier'
./publishSource:
./publish.GitHub forces us to use the master branch for the build output... so the actual source is in the branch
dev.Update the gems with:
bundle update
git add Gemfile.lock
git commit -m 'update gems'His website was originally written in markdown, however those were deprecated in favour of AsciiDoctor when Ciro saw the light, rationale shown at: markdown-style-guideuse-asciidoc
GitHub pages is chosen instead of a single page GitHub README.adoc for the following reasons:
- Ciro will want some unsupported extensions, notably mathematics, likely with KaTeX server side:
- github.com/asciidoctor/asciidoctor/pull/3338
- stackoverflow.com/questions/11256433/how-to-show-math-equations-in-general-githubs-markdownnot-githubs-blog
- g14n.info/2014/09/math-on-github-pages/
- stackoverflow.com/questions/11256433/how-to-show-math-equations-in-general-githubs-markdownnot-githubs-blog
- www.quora.com/How-can-I-combine-latex-and-markdown-in-GitHub
- when GitHub dies, Ciro's website URL still lives and retains the PageRank!
It would be really cool to have a PageRank-link algorithm that answers the key questions:However, Ciro has decided to leave this for phase two action plan, because it is impossible to tune such an algorithm if you have no users or test data.
- what is the best content for subject X.For example, if you are reading
cirosantilli/riemann-integraland it is crap, you would be able to click the buttonwhich leads you to the URL: ourbigbook.com/subject/mathematics. This URL then contains a list of all pages people have written about the subjectVersions by other authors
mathematics, sorted by some algorithm, containing for example:This URL would also contain a list of issues/comments that are related to the subject. - who knows the most about subject X. This can be found by visiting: ourbigbook.com/users/mathematics "Top Mathematics users", which would contain the list of users sorted by the algorithm:
Perhaps it is also worth looking into ExpertRank, they appear to do some kind of "expert in this area", but with clustering (unlike us, where the clustering would be more explicit).
Other dump of things worth looking into:
Stack Exchange solves to a good extent the use cases:points of view. It is a big open question if we can actually substantially improve it.
- I have a very specific question, type it on Google, find top answers
- I have an answer, and I put it here because it has a much greater chance of being found due to the larger PageRank than my personal web page will ever have
Major shortcoming are mentioned at idiotic Stack Overflow policies:
- Scope restrictions can lead to a lot of content deletion: closing questions as off-topicThis greatly discourages new users, who might still have added value to the project.On our website, anyone can post anything that is legal in a given country. No one can ever delete your content if it is legal, no matter their reputation.
- Although you can answer your own question, there's no way to write an organized multi-page book with Stack Exchange due to shortcomings such as no table of contents, 30k max chars on answer, huge risk of deletion due to "too broad"
- Absolutely no algorithmic attempt to overcome the fastest gun in the West problem (early answers have huge advantage over newer ones): meta.stackoverflow.com/questions/404535/closing-an-old-upvoted-question-as-duplicate-of-new-unvoted-questions/404567#404567
- Native reputation system:
- if the living ultimate God of
C++upvotes you, you get10reputation - if the first-day newb of
Javaupvotes you, you also get10reputation
- if the living ultimate God of
- Randomly split between sites like Stack Overflow vs Super User, with separate user reputations, but huge overlaps, and many questions that appears as dupes on both and never get merged.
- Possible edit wars, just like Wikipedia, but these are much less common since content ownership is much clearer than in Wikipedia however