Rosebush and moving over from Posterous

Yeah--I am now REALLY liking how Blogger works and operates in comparison to Posterous.

Yes, Posterous did offer the convenience of being able to email entries. However, emailed entries tended to have a large amount of issues in formating and tagging. They also had to be monkeyed with to be posted to the other networks. That is, an item would need to be sent to twitter@harlot.posererous.com to get posted to Twitter. facebook@* for Facebook, etc.

I dunno--a few glitches could have been worked out. It also does not help that I never had any followers, nor did I really see anybody that I would want to follow. In comparison to twitter, tumblr and facebook. I dunno--it just kind of lonely.

Yes, there were a few regular people who would view my posts. I dunno--I rarely got more than 15 regular unique hits.

Add the void of Posterous and the general glitchiness of the functionality--I kind of tried out a few Blogger posts... and wow... I am in love with blogging again.

Now then--Rosebush. The item that will combine and replace my social networks. My goal to make NIMHLabs able to pull large amount of data from various social networks I am on, and put it in one place.

Rosebush will also make it easier to run the various items I desire to post onto my site. Typically these will be Source Books that I either have permission to post--or nobody really is going to question me on them, as the systems are not used for various reasons.

It will also contain a Pokedex, information on Rockman.EXE, The Mother Series--and a fair amount of data on various other games I am interested in.

Now before you suggest I go with a Wiki, I will say no. Even if I lock it down, Wiki's generally have formatting issues to them. Generally it makes it very hard to pull information from one item, and cross reference to another. Automatically posting information onto one Wiki page, for stats, generally does nothing to affects stats on another page.

Lets take TvTropes for example. A wonderful wiki. The issue is the tropes do not cross reference with the series. Posting a trope onto a series, does not put that series as an entry on the trope. Putting a series as an example on a trope, does not update that trope into the series.

This is a simple example, however this sort of thing tends to grow bigger and bigger on these sites. Making it worse by require manually adding data from one page to another.

That is why I do not want a Wiki for NIMHLabs.

There should be a few spots where user editable content may be present. Perhaps at some point. But NIMHLabs really should not run a Wiki.

It really does not help, that the main example of Wiki software people spooge over has some of the worst layout issues to its name. This being MediaWiki.

Open up a page in MediaWiki, using most of the layouts available. A fine layout, right? Okay--what do you click on the page? Is it in a decent spot?

My fixes would be:

• Search bar at the top and left
• Login information next to it
• Links to edit, view history, watch or discuss the page moved to a shorter side bar. Login links go there. As would the "random page" link.
• All those other links you see mucking up that layout? All those links that most people rarely, if ever use? All those links that generally add the idea of "what am I clicking on?"? They get moved to the footer of the page.
• The license this page is under is moved to an easier to see spot on the page. If it is a CC page, move that information from the bottom of the page, to possibly the upper right.

I mean, yes--this could be fixed with a simple skin. However, this is layout. We also have to deal with update propagation and another fun thing:

Most wiki's do not have a tagging system to them. I mean, yes, a template can be added that identifies other pages that should be looked at. This however is a little bit too much for this to really be the best plan.

Look at the Boorus. Gelbooru, e621, Paheal.Net--heck look at a site known as the Overbooru. What has made them some of the best picture gallery sites out there? I mean--copious amounts of porn is not it. I mean--there are other systems that people do not bother with that provide this as well.

Tagging. Each image is tagged by the viewers with any tag that may arouse interest in the depiction, no matter how tasteless or crude the image depicted is. These tags help people find everything faster.

This kind of is part of what I want Rosebush to have. Pages will be tagged, and then items in the page will be linked in a form of what would cause the Wikiwalk effect in anybody viewing the site.

This is generally increased, when I plan to add various written works to the site. Mostly in articles, howtos, stories and dreams I have had to the site. I also want to add a copy of ESR's Jargon File onto NIMH Labs. The Jargon File will likely be user edited.

The subject matter, I want to get with this site, is, well, mostly stuff on graphic design, programming and literature. Subject matter designed to allow you to more critically look at the world around you. Be able to look at advertisements, point out stop value, presentation of need and fulfillment of need show in the ad. To be able to point out how ads work. Literature would mostly be to help you understand how English works. Why and how various stories are written like they do and are. Programming as a means of learning how computers work and why. This would likely require a fair amount of history on this site too. Naturally the Legal System would also be a requirement of NIMH Labs to be covered.

However, while I am doing this, I would need to demonstrate that I do know stuff here. That I am not just another Foghorn Leghorn blabbing on at the mouth on something.

Which is why, I need to program a few things, draw a few comics and write a few stories as I am doing this.

Also, as I go though doing the literature--my plans are to have it annotated. That is, various items could be rolled over, and a little dialog would show above it, noting a few annotated elements on that item. Possibly, with a set of cliff notes presented from annotations, should this get user edited.

The issue is--well... the basis of the project, right now is Pylons. Though--it is generally a pain for me to work with here. I mean--if I were to try to code this from scratch I would have a clue as to how to do it. However the actual coding would "take longer than if I did it with Pylons" (if I really could figure out Pylons), and "would benchmark much slower than Pylons" (which I am likely to agree there).

The issue here is it is the same issue with RoR, Zope, and Catalyst.

I mean--these "simple" scaffolding systems have gotten to the point of impeding development for me. I mean, yes, I run a single command and most of the files I need to do this are automatically generated. However, the amount of editing I would need to do anyways kind of defeats any gain by the auto-gen. I mean, even then, what is generated is most of the time, only template files with nothing of interest in them anyways. May as well be done via scrap.

It also does not help that, to be frank, documentation sucks on RoR and Pylons. I mean, if there was more to read on how to do anything with it--rather than a few simple scripts on Pylon's site, with links to other sites, and me generally having no idea on how the developers were thinking of doing stuff when they put the stuff all together (it was "lightly" touched on in the Pylon's docs).

And even Pylon's ruefully lack of documentation is a lot better than RoR. Which did not have any documentation on it what so ever. You'd think with a language that brought about rdoc, people would be a lot more anal about commenting their work. Generally every rdoc generated library of documentation mostly consists of the modules and function names. Very rarely is the purpose, use, functionality of any of the entries explained. It relies on me thinking like the programmer was at the time of making the library, and grepping through source files given for how each function was implemented.

Trust me: when this is how people learn, you get a lot of people using your library VERY wrong. Usually getting bad programs, that while they do work, they are generally fundamentally wrong in how they were designed to operate. Incidentally, most of the libraries that propagate this ruby phenomenon further tend to be fundamentally wrong in how the use their libraries and work and operate.

Ruby could have been a much better language. Really it could have. That is, if people actually documented how to use what they build in it better. This, unfortunately, has caused much reinvention of the wheel in Ruby.

Catalyst I would use--however I am still having issues with cpan properly installing it on Dreamhost. Pylon only got used due to Python's ability to do "easy_install" to grab something from cheese shop, without whatever I am installing to screw up for some reason. Usually the fixes for these screw ups tend to be, "something is wrong with your cpan configuration. Fix it and then ask me again", "I dunno" and "oh, it should just install, if you type in the command"

I have complained in the past about cpan generally functioning horridly. I have tried it again--and well... if functions slightly better. Kind of like how a chimpanzee with a lobotomy does function better than a comatose patient with an IQ less than a tomato.

I will however, begin work on Jerl 6 soon. Provided that nobody who actually has competence has you know, done their job rather than wait for me to do something.

I dunno, projects I currently have on my slate for 2010:

• Rosebush & NIMH Labs
• KidRadd.Org
• Street Slifer
• SSBZ (this has been upgraded to being done by the end of 2010)
• libNIMH
• Kid Norrin
• Mofo Online
• P*DA ZX
• Geist
• Lamia Browser

*sigh*

It has been a real kick out of me being out of commission for the first three months of the year, due to hypothermia. Though, I am back on my feet, and kicking.

Here we go!