This page is about the pending Noosphere rewrite. Some related discussion (mainly of "nitty gritty details") is already taking place at Noösphere Niggles.
I would like it if all of the ideas we've been kicking around on the wiki are considered. The rewrite could come at a great time – the perfect chance for us to "Apollonize" our "Dionysian" ideas.
In particular… I hope serious consideration will be given to:
as well as all of the relevant Feature Requests. Even the idea about [[thoughts on the viability of free-software style business|for-pay question answering]] should be considered. It may be that some or all of these features should be developed, but not as part of Noosphere proper, rather, as modules that will fit into Noosphere. In this case, good support for outside modules should be built into the system. --jcorneli Sat Apr 09 15:57:57 2005 UTC
Has anybody used Ruby on Rails, http://www.rubyonrails.org . I'm going to look into it, no offense against python or perl, but I think they are not modular or extensible enough. Reuse Reuse … Reuse …
--bloftin
Personally, I have no idea, but I think that people in the know should update the Plans section of the wiki. --jcorneli Sat Apr 09 15:57:55 2005 UTC
To me, a more serious problem than planning a timeline is finding people to work on this project. As we all know all to well, there are only a handful of us who are involved in the activities discussed here, and only a subset of that handful has contributed to Noosphere. Moreover, all of us seem rather busy with a variety of activities, so we really need to interest more people in participating if a project of the sort outlined above is to be realistic in the foreseeable future.
This topic of needing to get more people aboard has come up before. I have been thinking about it, but still do not have a better answer than to keep thinking about it and, in the meanwhile, continue with what we have been doing to publicize our activities. However, in this case, we maybe should think of Planet Math contributors as a potential contributors to the code as well. To make this feasible, I think it would also help to put some thought and effort into making it easier for people to contribute. In particular, I have in mind trying to make things easier for people who are not that well-versed in programming, specifically PERL programming.
One thing that would help and not require much effort is providing interested people with background information. Surely, there are a lot of good textbooks, manuals, tutorials, faq's, and the like about PERL out there; it would be nice to have an short annotated bibliography to which we could refer the interested beginner. Likewise, I suppose that there is some other background material which might be necessary. I furthermore suppose that much of that material is to be found in Aaron's thesis and the research articles referenced therein, so it might suffice to simply add these to the bibliography.
Another thing that would help is to make hacking noosphere as user friendly as possible. As I see it, if one can lower the learning curve, the probability of getting users (i.e. Planet Math contributors) to contribute to the code as well goes up. If the laws of physics apply here, the probability of penetrating the barrier depends on the barrier height exponentially, so even small improvements can effect appreciable improvements.
Also, I think that literate programming is a great way of making programming more user-friendly. By explaining what is being done in humanese as well as in computerese, it becomes much easier for people to follow the listing of a program and hence more likely for them to become interested in contributing to be program or even pointing out improvements like: "Wait a moment here! Why are you wasting valuable CPU time here initializing these variables when they are going to be initialized anyway by the routine which first uses them?" As Knuth once put it, TeX the Program can be used as enjoyable bedtime reading. Obviously, this is only true for technically minded readers, but these are exactly the sort of people who are interested in Planet Math.
Ultimately, I would like to see noosphere become like emacs in that it can document its own code and that one can use noosphere to write noosphere. Naturally, this is all a long term goal. Adding the documentation necessarry to make the program literate is a lot of work, so to some extent, this is a chicken-and-egg dilemma. I think that the way out is to do the easily done pieces first, to consider hacks and shortcuts that will allow one to have a reasonable approxiamtion to the intended functionality in the meanwhile, and to perhaps put "making programming noosphere more user friendly" high on the list of priorities in the hope that, when we get to the implementing the extra features, there will be more uf us to do the implementing. --rspuzio 9 April 2005
People (logan) actually working on the project can of course do whatever they want… but after taking a look at the trac page (see link at Noosphere) I would like to ask, unless there are pressing reasons to do otherwise, that the discussion components take place here on AM. If it turns out that the discussion is too overwhelming, then maybe we should consider moving it elsewhere. But for now, it seems like we could use more discussion here, and it also seems like the various ideas that we're kicking around on AM will help inspire work on Neosphere. --jcorneli Wed Apr 13 05:13:29 2005 UTC
At the PM summit, Aaron said he'd ping Logan to see what the status of this project is looking like from its (candidate) lead developer's point of view. Any result on that, Aaron (or Logan?)?
Not to go around passing the buck, but Ocat's motion for a feasibility study for the HDM project makes me think that it would also be nice to have a full-fledged design document for the Noosphere rewrite. I'd be happy to contribute to this, but we'd probably need a pretty active developer team to make that a well-timed contribution. --jcorneli
Such a feasibility study might prove useful to show to potential contributors when we go door-to-door asking for donations. --rspuzio
Guess what — that time is arrived! --rspuzio 5 March 2006
What do you mean? --jcorneli
The current dicusions surrounding the whitepeper — now we are getting around to the business of writing a document to show contributors when we go knocking from door to door as opposed to simply talking with Ocat about doing something like this. —rspuzio
This is simply a page that talks about the implementation of the Slashcode in hope of learning something for the rewrite.
Now that, thanks to Joe's hard work over the past few months, the scholium system is a reality, I think we need to consider the relation of noosphere to arxana. In a sense, one could even consider arxana as a noosphere rewrite since a main reason Joe wrote it was to have an improved collaboration platform which did not have some of the shortcomings of noosphere and he was heavily influenced by Noosphere and PlanetMath in the design of Arxana.
At the very least, it would be nice to have compatibility — for instance, the Noosphere database could be designed in a way which makes it easy to translate it into article quintuplets and vice-versa. Also, scholium-based thinking could inform the rewrite.
However, I could also see the connection happen in more intimate ways as well. As I see it, the main strength of the scholium system is it's generality and flexibility — for instance, it would be easy enough to set up a semantic equivalent of PM or AM in the scholium system and even do such things as turns forum posts into entries with ease. However, arxana does not have the nice, user-friendly web interface which noosphere has. This suggests to me that one possibility would be to consider using the scholium system as a framework which takes care of organizing the various entries, posts and what not. This way, both projects might benefit — noosphere would gain a highly flexible, sophisticated system for combining documents and arxana would gain a user-friendly interface.
I suppose that more on this will be figured out when Joe and Aaron meet, but I thought it would be worth at least raising this as a suggestion. --rspuzio
I hope we can do something like you suggest; at least make the scholium system an alternative interface to the PM database. One problem with Noosphere (as I see it) is that the different layers of abstraction have tended to get blended together in one big block. (See levels of abstraction in PlanetMath.) One of my goals for the Noosphere documentation project is to unpack these, so I can figure out how I (and other wannabe hackers) can interface with the system. --jcorneli
One of the more advanced features of noosphere is the autolinker. It is also a feature which needs more work, as evidenced by all the complaints about wrong links and this problem only promises to get worse as PM grows.
I think that improving the autolinker is another area in which we can have a useul symbiosis between PM and HDM. While there are some obvious things that could be done to improve the autolinker, I also believe that there is only so much that can be done before one needs to use techniques of natural language processing to make the autolinker behave more intelligently.
As it turns out, we in HDM are very interested in the business of understanding mathematical liunguistics although we have not done much on this subject yet because we would like to have a linguist on board. Since basic researcher into the linguistics of mathematical prose will prove valuable to improving the autolinker as well, maybe we should consider some sort of collaboration here in the future. --rspuzio
[[Logan?]]'s writeup on http://planetx.cc.vt.edu/AsteroidMeta/No%c3%b6sphereProtocols is very good (mentioned at Noösphere Niggles). Sounds like he knows his stuff. Taking system design to that level gives it a very mathematical flavor, and given that PM and the Noosphere already exist, a rewrite using the most advanced techniques will be like cheese to mice – irresistable for those who are fascinated by computers and programming. A rewrite of an existing system is harder than starting from scratch, in some ways – the existing data must be converted and service interruptions are a major risk factor. But the nice thing about total rewrites is that the target is well understood and tangible, which helps keeps the effort well grounded. If PM can arrange this effort so that it becomes irresistable to students of computer science, it will almost write itself, at small cost – and those who successfully develop the rewrite along the lines described by Logan will find themselves with solid 6 figure employment prospects, or at the very least, solid resume line items and gold-plated references. P.S. IBM bills out people who can do what Logan is talking about and create a modern system at $300 per hour. And that isn't just a few hours here and there, like a plumber; we're talking 8 hours a day, 5 days a week for longggg time. The ROT is take the hourly rate and multiply by 2 to get the annual amount, net of holidays (which of course, IBM employees take.) So the incentive is there. [Unable to write template]100K. But money is not the issue for people who have a heart for the work, and it shouldn't be. The money comes in due course – money is society's way of saying thanks. --ocat 6-Mar-2006