Ake's corner

LIESMICH.TXT

This site mostly wasn't intended to be used as blog, but rather as playground/archive. So better go to stuff section and checkout some javascript demos instead.

Also, you can go to old index

NEOS UPDATES

Uploaded cutting-edge version of neOS webtop, it has some apps and features added, slightly updated look, legacy apps removed and code refactored. Test it

OVERDICTIVENESS

Created demo that generates new words in English or Russian, using spellcheck dictionary as input data and something like Markov chain model.

Check out

20200325

MICROKERNELED WEB DESKTOP

Finally uploaded P4 - web desktop without apps, but with API to connect external apps and use low-level platform features (a la syscalls) and shared storage. Well, part of things I've just mentioned isn't there yet as well as proper docs or sample apps.

20200317

WE NEED MOAR WIKI ENGINES

As I mentioned kiwi in previous, I uploaded it. You check it out here. This is something close to wiki and client-side static site generator. All its engine is one page that display wiki pages, which are just text files with some basic markup.

But when you started to check boundaries of posibilities and relevance of «unhosted» web apps, it's hard to stop. So I present you Teed, personal web notebook. It can be called stripped down version of Tiddlywiki as it follows its ideas, particularly encapsulation of content in page and storing changes via producing page replica with updates injected. However it inherits almost nothing of wiki and non-linear merits of Tiddlywiki, Teed (as its name says) is more like linear feed of unnamed card/notes/name it. Initially I think of it as storage for notes/links/todos etc, but it can be used as well as very simple blog with WYSIWYGy macroeditor (where WYSIWYGiness is more related to posts' overall layout than posts' content).

20200220

‹/NOSCRIPT›

I finally uploaded «next generation» of my one page sites. It is simply called sitejs and it shares ideas with similar projects I did before. There is index in javascript file, describing tree-like site structure that can contain data explicitly or just reference to it, so content will be loaded dynamically. Kinda new concept is «folder»-type pages that have multiple subpages and represent them. To some extent it was possible in kiwi (friendly reminder to me - reupload it), but you have to write index manually and naturally kiwi have flat structure. More fresh idea is to use this engine as hub for heterogeneous and distributed sources, i.e. you can add pages in index, refer html pages, json files that contains pages and so on.

You can see it here and doc drafts which describes underlying stuff a bit.

20200218

HOW DO I SHOOTDO WEB

So, even nobody cares, let me say what am I going to create and put here. I accept that idea of making consistent site was screwed up, at least for me. Most likely some sections will remain in old design, if it's fit (e.g. projects/stuff section at least for now), but with updated navigation and links, some will be updated to match new design. Keeping all pages synchronized and consistent (i.e. navigation, updates in uniform parts) without some generator (and I think my attempt to use generator routinely was failed) requires too much effort.

One thing I've wanted to add before is some kind of feed in parallel to main blog, but I don't see where can it be placed so it won't create clutter. I think I can make «ticker», but it's something [not so] slightly different. Another idea, that distantly related to previous one, is to create separate desktop-only version of site; main problem is that likely it will cause duplicating all content from main one. Other ideas about site is to create section or version in my native language, introduce some weekly post (I've even think it can be called ‹Blunt Corner› or ‹Ake's Blunt Corner› (ABC))

Another category is stuff that already done, but I'm still too lazy (and a bit of perfectionist) to upload it here and finish then. Another portion of old demos to Squirrel Scrapheap, another Webtop-like thingy, stuff from 2019 site, VCON project, pixelman and konquest TBS versions, client-side static site generator.

20210212

SURFACE SCRATCHING DIGITAL ARCHAEOLOGY

When I was burrowing through my archives and found older version of my site on neocities, it was funny to find that hyperlink webring widget is still functioning, despite it was abandoned, has most of its core team members (I think) left neocities and was squatted/restarted as completely different project multiple times. So I've decided to explore this "ancient artifact". It's easy to figure that if you have to pass you neocities username in widget URL, you can also pass any member's username to get their widget, to find other members' names you can just examine next and previous links.

So here is graph showing connections in Hyperlink webring that I was able to reconstruct with subrings showed as colours. Yes, due to subrings, it's large and complicated graph despite having only 26 nodes. Single node loops is subrings that have only one member.

Ah, and said widget

20210212

WAIT, I'VE SEEN IT IN 2018

Yes, I've decided to revive (I think) site and rewind to it's old form. Golden era (ha), you know, once its look was even copied, so I think it's something valuable and kinda reflecting my personality (yes, this generic faceless thing). BTW it's slightly differs from those variants and probably will continue evolving.

Most of redesign is still in work and a lot of pages are disconnected at the moment, but if I haven't started it, it would last on disk for years. To browse site you may go to previous version.

20210210