NUMA going forward

[Fraxtil]

yeah, i think Gforce should take over numa, having too seperate archives would really suck

As previously stated, I have no experience with the Google App Engine and little experience with Python. Learning all that plus Numa's infrastructure would be much more of a hassle than it's worth.

[TheSeer]

Poking around, the verification email can't be sent from the default codebase (no idea why) - for the purposes of testing I just bypassed it completely via setting a manual EmailHash (I guess I could've just told it to tell me :o) and going to the verification link.

Edit: Removed a bit. Turns out it's just me not knowing how the whole map ecosystem works.

[Tunco]

This'll be great. At least some change will be good. Very excited for it.

[jean-luc]

I'm still willing to help whereever I can, but unfortunately I've been very busy lately and learning Python and GAE has simply not been in the cards (Robotics competition season, Speech & Debate, advanced classes, programming contest, SAT/ACT, etc..., and my free time has mostly been spent working on a commercial web app (http://scholaro.us) which is on Rails, and I can't quite learn two different web frameworks at once.)

[Amadeus]

That was really insightful. Thanks, Arachnid.

[LittleViking]

I'd really love to work on this, but I just can't get my local site set up. We have database info in CSV form but no easy way to get that into App Engine's datastore. Anyone have any luck setting up the datastore at all?

[Nexx]

Perhaps this is not a good suggestion or not the place for it, but I'm going to throw it out there anyway: should NUMA go through another language change? It seems to me (and SoY, in that other thread) that GoogleApps-powered Python is not as well represented by coders in this community as other languages (like good ol' PHP). Perhaps it would greatly ease the passing of the reins if the site was remade again back in PHP, or in another appropriate language. But then maybe that'd be equivalent to just helping out GForce to get up AltArc more quickly and completely, and seeing if NUMA data can be transferred or something, I don't know. But this seems to be the main issue with why the reins haven't been passed already, so I'd like to spark some discussion on it. On the other hand, if it seems like LV is destined to fill Arachnid's spot, maybe there need not be discussion.

[Amadeus]

Perhaps this is not a good suggestion or not the place for it, but I'm going to throw it out there anyway: should NUMA go through another language change? It seems to me (and SoY, in that other thread) that GoogleApps-powered Python is not as well represented by coders in this community as other languages (like good ol' PHP). Perhaps it would greatly ease the passing of the reins if the site was remade again back in PHP, or in another appropriate language. But then maybe that'd be equivalent to just helping out GForce to get up AltArc more quickly and completely, and seeing if NUMA data can be transferred or something, I don't know. But this seems to be the main issue with why the reins haven't been passed already, so I'd like to spark some discussion on it. On the other hand, if it seems like LV is destined to fill Arachnid's spot, maybe there need not be discussion.

I think it just may be worth it, because it might be better to completely redo Numa rather than fix it, as using Python, it is going to be so difficult. If we can get Numa onto a different coding, then more people would be able to contribute and participate in helping out Numa.

[chocollama]

I tink that NUMA should have the ability for you to see who is online and make friends and rivals. Get back to me on what you think of my ideas.

[Nexx]

I think it just may be worth it, because it might be better to completely redo Numa rather than fix it, as using Python, it is going to be so difficult. If we can get Numa onto a different coding, then more people would be able to contribute and participate in helping out Numa.

Actually, a huge reason why I made my original post (though I forgot about this at the time I made the post) is because Python on Google Apps:
(1) can't overcome the results truncated problem, and
(2) can't sort listings!

Sorting map listings is key to navigation! It's basically the backbone of Gforce's AltArc, and it works great for sites like DeviantArt and whatnot. It'd be perfect. The only problem is the Python with Google Apps can't do it (or can't do it well). And that sucks. That means that instead of recent top-rated, where you have THE COMMUNITY picking out what's good just by doing what it would do anyway, you have to result to other systems like bitesizing or float/sink to highlight recent good maps. It takes work to implement and maintain those systems, and they're not as efficient as a recent top-rated anyway. It seems to me like it could all be resolved if the site was restructured in another language.

[LittleViking]

I think it just may be worth it, because it might be better to completely redo Numa rather than fix it, as using Python, it is going to be so difficult. If we can get Numa onto a different coding, then more people would be able to contribute and participate in helping out Numa.

Actually, a huge reason why I made my original post (though I forgot about this at the time I made the post) is because Python on Google Apps:
(1) can't overcome the results truncated problem, and
(2) can't sort listings!

Sorting map listings is key to navigation! It's basically the backbone of Gforce's AltArc, and it works great for sites like DeviantArt and whatnot. It'd be perfect. The only problem is the Python with Google Apps can't do it (or can't do it well). And that sucks. That means that instead of recent top-rated, where you have THE COMMUNITY picking out what's good just by doing what it would do anyway, you have to result to other systems like bitesizing or float/sink to highlight recent good maps. It takes work to implement and maintain those systems, and they're not as efficient as a recent top-rated anyway. It seems to me like it could all be resolved if the site was restructured in another language.

GAE can definitely sort listings, and it's really ideal for a recent top-rated page as opposed to an all-time top-rated. Since App Engine can only query 1000 results at a time (1-2 weeks of Numa maps), it would only ever give the most recent maps without multiple calls. If the rating column on the maps table was indexed (easy to do), a recent top-rated page could be added right away.

The reason it's not there is that Arachnid and other people believe the rating system of old Numa was fundamentally flawed. The ratings at the top were so close, a single vote lower than 5 could be enough to push them down. Maps jumped around all the time. The ratings difference between the number 1 map and the number 15 map might have been a few hundredths of a point.

As for changing languages, I find the Python system really clean and nice. The only problem I'm having is getting the sample data imported. Once that's passed, the rest of development should be peachy.

[Nexx]

GAE can definitely sort listings...

Yes, well, it can. But Arachnid's response to such a suggestion was:

"Implementing multiple sort orders entails an exponential increase in the number of indexes required - not really practical - and is only used by a very few people.

Future improvements will focus on improving the existing filtering and sort, instead, hopefully to the point where everyone's satisfied with it."

And top-rated represents only one way of sorting. Sorting by # of rates would be very useful as well (and, ideally, also by # of faves).

The reason it's not there is that Arachnid and other people believe the rating system of old Numa was fundamentally flawed. The ratings at the top were so close, a single vote lower than 5 could be enough to push them down. Maps jumped around all the time. The ratings difference between the number 1 map and the number 15 map might have been a few hundredths of a point.

It seems like you're jumping back and forth between two topics here. There's the topic of an all-time top-rated and that of a recent top-rated. In old NUMA, they were one and the same. As a result things jumped around all the time. (Actually, various people argued that it was dumb either because it moved too much or because it didn't move at all - a lose-lose situation for those who enjoyed the top-rated page.) But even more than that, once a recent map got, for example, five 5's, it would kind of become this hyperinflated rating contest where those who liked the map would give it 5's to keep it afloat because if they did anything less, it'd get knocked way back. There were also the people who went "this doesn't belong here" and would smack it down just to smack it down. All of this is a problem of the structure of top-rated as implemented in old NUMA (not a problem with top-rated itself). But yes, if both recent and all-time listings existed side-by-side, something would have to be done so that a new map with five 5's doesn't make it onto top-rated of all time (unless it gets like 40 5's, that'd be another matter altogether).

The other issue is true of even a good top-rated system: ratings will be so close that it doesn't mean anything. I still stick by my idea of emulating DeviantArt. In fact, I might even suggest that a top-rated of all time doesn't give a 1-10 listing but instead shows only thumbs in a manner similar to active authors. If you start numbering the top-rated maps, bad things happen.

[Arachnid]

I think it just may be worth it, because it might be better to completely redo Numa rather than fix it, as using Python, it is going to be so difficult. If we can get Numa onto a different coding, then more people would be able to contribute and participate in helping out Numa.

Actually, a huge reason why I made my original post (though I forgot about this at the time I made the post) is because Python on Google Apps:
(1) can't overcome the results truncated problem, and
(2) can't sort listings!

Sorting map listings is key to navigation! It's basically the backbone of Gforce's AltArc, and it works great for sites like DeviantArt and whatnot. It'd be perfect. The only problem is the Python with Google Apps can't do it (or can't do it well). And that sucks. That means that instead of recent top-rated, where you have THE COMMUNITY picking out what's good just by doing what it would do anyway, you have to result to other systems like bitesizing or float/sink to highlight recent good maps. It takes work to implement and maintain those systems, and they're not as efficient as a recent top-rated anyway. It seems to me like it could all be resolved if the site was restructured in another language.

GAE can definitely sort listings, and it's really ideal for a recent top-rated page as opposed to an all-time top-rated. Since App Engine can only query 1000 results at a time (1-2 weeks of Numa maps), it would only ever give the most recent maps without multiple calls. If the rating column on the maps table was indexed (easy to do), a recent top-rated page could be added right away.

The reason it's not there is that Arachnid and other people believe the rating system of old Numa was fundamentally flawed. The ratings at the top were so close, a single vote lower than 5 could be enough to push them down. Maps jumped around all the time. The ratings difference between the number 1 map and the number 15 map might have been a few hundredths of a point.

As for changing languages, I find the Python system really clean and nice. The only problem I'm having is getting the sample data imported. Once that's passed, the rest of development should be peachy.

Exactly right. As far as importing data goes - have you looked at the GAE articles on data import, such as http://code.google.com/appengine/docs/p ... gdata.html ? I used the old bulk loading system, but the definition files should be in the codebase somewhere. Failing that, you can just insert a few sample records of your own - you shouldn't need much data to do development. :)

[LittleViking]

As far as importing data goes - have you looked at the GAE articles on data import, such as http://code.google.com/appengine/docs/p ... gdata.html ? I used the old bulk loading system, but the definition files should be in the codebase somewhere. Failing that, you can just insert a few sample records of your own - you shouldn't need much data to do development. :)

I did read through that article, but I just kept running into problems trying to import. I might just add some sample records of my own like you said, though.

[im_bad_at_n]

I'm currently learning basic html and xhtml coding. Next to learn is CSS. I don't think that helps though.

I'm all up for redoing NUMA completly. Hoewver before main coding begins, we need a thread asking the general community what they want to see, for example, do we want a "top favorited" section? And all questions like that. Then from those results, we could determine the best coding type (python, etc.) and from there the active coders. Just a suggestion. Might be a bad one =/

[origami_alligator]

I'm currently learning basic html and xhtml coding. Next to learn is CSS. I don't think that helps though.

I'm all up for redoing NUMA completly. Hoewver before main coding begins, we need a thread asking the general community what they want to see, for example, do we want a "top favorited" section? And all questions like that. Then from those results, we could determine the best coding type (python, etc.) and from there the active coders. Just a suggestion. Might be a bad one =/

Why make a thread? The most popular suggestions are HERE.

Also, you need to learn Python and Java and become very adept at coding in those languages if you hope to take over NUMA.

[Amadeus]

With all due respect, I'd rather have it handed over to someone far more involved in the community. I feel like it would be too easy to lose contact with Unreality and have the entire thing out of our hands.

Agreed. Although Arachnid did a great job starting off the site, we need someone who will be there consistently, so we don't have to wait months or years to get bugs fixed.

[Fraxtil]

I'm currently learning basic html and xhtml coding. Next to learn is CSS. I don't think that helps though.

I'm all up for redoing NUMA completly. Hoewver before main coding begins, we need a thread asking the general community what they want to see, for example, do we want a "top favorited" section? And all questions like that. Then from those results, we could determine the best coding type (python, etc.) and from there the active coders. Just a suggestion. Might be a bad one =/

If you're thinking of doing that, you might as well start with AltArc, which implements many features that were lost in the transition from PHP to App Engine, along with others that were never in NUMA to begin with. I think I might want to clean up some of my code before I open-source it, though...

[Destiny]

Is AltArt still in the works? I've not seen any change since i last checked it out six months ago.