Marathon Donation App - Call for testers

2012-01-11 06:55:04 pm

Quote from Emptyeye:

By censoring, I guess you mean comments and not names of donators?

By censored, he names in the website currently are 'censored' as "first name, last initial", though I'm not sure if that's absolutely necessary, since we've been reading off full names since the beginning. Maybe we should have a per-donor flag that indicates whether a given donor would rather have their full name, abbreviated name, alias, or just their ID listed, and then have it default to the abbreviated name? (People could specify their preference in their comments.)

At the moment, there's no censoring of individual words in comments, just inappropriate comments are (hopefully) listed as 'denied'.

Itsu de Mo

2012-01-11 07:08:33 pm

Arrested for felony abuse of emotes

I was really pleased how relatively easy it was to pick up on the hidden side (like assigning bids and such). Cool Matty literally taught me how to run it and 5 mins later, I was absolutely fine on it. ^^

Any other issues I had with it were already mentioned (by SMK more or less).

dballin

Edit history:

dballin: 2012-01-12 05:33:50 pm

2012-01-12 05:31:45 pm

Balls jerky

Any issues I had, you posted. Most likely because they were discussed on site, but here's another one I thought of: Would it be possible to add a flag that lets the person reading donation comments know to treat it as anonymous? I'd say like 99% of the time, the person entering bid data got to the donation first so they could flag it and prevent "we've got a $10 donation from John Smith who says 'please don't read my name' ...whoops"

I think it's also important to note, that the program worked quite well and was incredibly easy to use once you figured out the quirks. Don't be too hard on yourself about any short-comings especially on something you did in your spare time and for free.

Cool Matty

2012-01-12 09:16:40 pm

Crawlathon WR, get down on my level.

There was a ton of minor things that were requested as features. I think you nailed most of them, and I'm too sick with this stupid marathon-flu to think of others, except the ability to counter-donate on a challenge (aka negative donation). In the end though I think a lot of the problems we had came down to simply an underestimate of the ridiculous number of donations we were getting, and so the backend wasn't always able to sustain it. Not just with UA's site, but chip in as well.

And yes, whenever we get settled and our lives return to normal, UA & SMK, I will gladly host the DB and site stuff on my dedicated server.

SMK

2012-02-14 06:14:05 pm

*Sigh*

So, I took a look at the code the other day, and shuddered. I know I promised all sorts of updates/upgrades, and SGDQ is coming up fast, but its really gotten to the point where I can't really maintain this program in its current state, and I'd need to take some time to re-structure it to make it more maintainable.

So.

I'm putting this forward as an open call to anyone who'd be willing to help me work on the application. I know this has been my baby for the last year and a bit, but I'm not really attached to anything in it and I'd be willing to take on any suggestions or preferences that other developers would have. I wouldn't really be expecting a large commitment or anything.

Poxnor

2012-04-10 10:12:58 am

Moo! Flap! Hug!

I honestly don't know if this is a feature anyone would want, so please don't take this as a request. I'm putting this forward as an idea.

When I've done raffles in previous (non-SDA) marathons, I've generated a verification ticket. Essentially, the program takes the current time (millisecond precision) as the seed to the PRNG to select a winner. Then, it outputs the winner as well as the time (ms from Unix epoch, and translated to UTC) and the list of entrants. Then, anyone can re-input the number of entrants and the time to the program (in "verification" mode) to verify that the raffle was conducted in an unbiased manner. I did raffles like this just to help dispel any false notion that the raffles were held in some biased or unfair fashion (true, the raffle still could be biased by fracking with the seed, since I was using millisecond precision, but it's the best I could do).

If you'd like, I could either send the code for that (it's a simple console program) or try myself to integrate it. That said, I'm a total crypto-nerd, and this could be totally superfluous Smiley

I honestly don't know if this is a feature anyone would want, so please don't take this as a request.  I'm putting this forward as an idea.

When I've done raffles in previous (non-SDA) marathons, I've generated a verification ticket.  Essentially, the program takes the current time (millisecond precision) as the seed to the PRNG to select a winner.  Then, it outputs the winner as well as the time (ms from Unix epoch, and translated to UTC) and the list of entrants.  Then, anyone can re-input the number of entrants and the time to the program (in "verification" mode) to verify that the raffle was conducted in an unbiased manner.  I did raffles like this just to help dispel any false notion that the raffles were held in some biased or unfair fashion (true, the raffle still could be biased by fracking with the seed, since I was using millisecond precision, but it's the best I could do).

If you'd like, I could either send the code for that (it's a simple console program) or try myself to integrate it.  That said, I'm a total [url="http://xkcd.com/538/"]crypto-nerd[/url], and this could be totally superfluous :)

CyberBotX

2012-04-14 10:51:21 pm

Insane Killer Robot

I know I already talked to you about this on IRC, SMK, but I'll post it here anyways.

It's too late for SGDQ, but after the marathon, I'm offering to help with a rewrite in C#. While this won't make it as easy to use on anything non-Windows (but still possible if compiled with Mono on another OS or just run through Wine), I don't see that being a big issue, and even though I lack good knowledge of Java, I think C# is a bit better of a language (despite it being Microsoft's).

SMK

2012-04-17 08:49:10 am

I should probably check the forums more than once a week.

@Poxnor: It's not a bad idea, and wouldn't be too hard to implement (add the drawing seed as a column to the prize entity, and then have an option to use that instead of generate a new one), but its pretty low priority, considering I haven't heard any complaints about prizes seeming unfair. I'll take a look if I have spare time.

@CBX: Yay! (maybe we can convince Dessyreqt to help too). The non-windows thing is a non-problem imo since the only non-windows box we had to deal w/ in the past was tri-hex's mac, which had a virtual windows on it anyways. Also, what version of visual studio are you using? I only have the full version for 2008, and the express 2010 one I have is only configured for XNA atm, so I'd have to change it over to winforms or WPF or something in such a case.

CyberBotX

2012-04-17 09:10:59 am

Insane Killer Robot

Well, at work we have VS 2010 on our development machine, but I am not going to hold myself to that because it's company owned. At home I have a copy of VS 2010 that I obtained (take that how you will), but I do think VS 2010 Express could do the job regardless.

Poxnor

2012-04-18 05:15:44 pm

Moo! Flap! Hug!

Quote from SMK:

[...]but its pretty low priority, considering I haven't heard any complaints about prizes seeming unfair. I'll take a look if I have spare time.

Yeah, that's what I figured. I haven't heard a single complaint. It just came up in casual conversation in my chat one evening, and Mike suggested that I mention it here to see if there was any interest. I figured there wouldn't really be Smiley

SMK

Edit history:

SMK: 2012-05-08 09:38:09 pm

2012-05-08 09:37:52 pm

New release is up at the usual location. As usual, its abysmally untested, so a few bugs may have crept in when I was fixing the other ones. I'll try testing it a bit on the weekend when I have a chance, but the main parts should be working fine.

UraniumAnchor

2012-05-08 09:47:15 pm

Not a walrus

There haven't been any database schema changes, right? I can set up a new database with the same structure and have it work?

SMK

Edit history:

SMK: 2012-05-08 09:58:35 pm

2012-05-08 09:58:19 pm

Nope (or yep, I guess), its all the same as before. And you should just be able to pull in all the run information using the google spreadsheet function in the program too. The only thing you'll have to add manually is prizes.

Eternalspirit

2012-05-11 01:08:21 pm

Meh.. I'll just officially add myself to the "help with converting to C#" group. Hopefully I can be of service.

mikwuyma

2012-06-09 10:59:35 pm

My feelings on The Demon Rush

I really should have made this post after the marathon so I could remember more, but more important stuff like all you can eat sushi, Fun Unlimited, Anarchy Reigns demo, and Rayman Origins got in the way.

The big thing that needs to be worked on in this program is optimization. Refreshes and saves taking forever is the #1 reason why donation processing takes so long IMO. UA did a full merge on his tracker website and it took 30 seconds. I started a full merge at the beginning of FF6 100%, and it still wasn't done by the end of the marathon. Sad

I'm guessing converting this program to C# would help a lot with the optimization.

Another thing is that some of the safeguards and preventions that were added for AGDQ were either taken out or glitched during SGDQ. I'm talking about double-applying bids (occasionally me and DW would apply the same donation to a challenge at nearly the same time, resulting in double the amount it should have), and filtering out any donations that were less than a dollar (there weren't many, but I saw a couple).

One much more minor thing is that I think the program should auto-refresh after you save information so you can actually see your changes.

BTW, SMK, I'm not shitting on the program, because it is a great program and does make bid processing 10x more efficient, but if we want to handle the high volumes of donation traffic we will receive during the next AGDQ, then the program needs some work.

I'm sure there's some other stuff with the program, but I can't remember it off the top of my head.

I really should have made this post after the marathon so I could remember more, but more important stuff like all you can eat sushi, Fun Unlimited, Anarchy Reigns demo, and Rayman Origins got in the way.

The big thing that needs to be worked on in this program is optimization. Refreshes and saves taking forever is the #1 reason why donation processing takes so long IMO. UA did a full merge on his tracker website and it took 30 seconds. I started a full merge at the beginning of FF6 100%, and it still wasn't done by the end of the marathon. :(

I'm guessing converting this program to C# would help a lot with the optimization.

Another thing is that some of the safeguards and preventions that were added for AGDQ were either taken out or glitched during SGDQ. I'm talking about double-applying bids (occasionally me and DW would apply the same donation to a challenge at nearly the same time, resulting in double the amount it should have), and filtering out any donations that were less than a dollar (there weren't many, but I saw a couple).

One much more minor thing is that I think the program should auto-refresh after you save information so you can actually see your changes.

BTW, SMK, I'm not shitting on the program, because it is a great program and does make bid processing 10x more efficient, but if we want to handle the high volumes of donation traffic we will receive during the next AGDQ, then the program needs some work.

I'm sure there's some other stuff with the program, but I can't remember it off the top of my head.

DW

2012-06-09 11:18:41 pm

One issue we ran across, along with the double assigning of bids, is the prize drawing had really weird default times to it, so it was hard to assign a time. Part of the problem was the time zone issue, which I guess isn't a problem at AGDQ, but it is at SGDQ.

But yeah, the program is fairly easy to use, even for someone like me, who knows this stuff just enough to make trouble. I had a little difficulty finding bids and especially prizes, particularly specific perlers. Searching prizes by game didn't work well, from what I remember.

mikwuyma

Edit history:

mikwuyma: 2012-06-09 11:58:57 pm

2012-06-09 11:56:09 pm

My feelings on The Demon Rush

Ohhhh yeah, that's another thing. Searching prizes, challenges, and bidding wars by game name was incredibly awkward. Why isn't there an option to just type in the game name? You had to know part of the prize, challenge or bid name to efficiently find it.

UraniumAnchor

2012-06-10 01:11:41 am

Not a walrus

Java vs C# isn't going to make a big difference in the efficiency, the issue is that whatever database library we were using is either poorly optimized for the sort of traffic we were throwing at it, or we were using it incorrectly. Switching how the program interacts with the website will be a big help, so that shouldn't be an issue any more.

SMK

2012-06-10 09:06:48 am

Quote from UraniumAnchor:

Java vs C# isn't going to make a big difference in the efficiency, the issue is that whatever database library we were using is either poorly optimized for the sort of traffic we were throwing at it, or we were using it incorrectly. Switching how the program interacts with the website will be a big help, so that shouldn't be an issue any more.

Indeed. The bottom line is that the program wasn't even designed to interact with a remote database in the first place, and I really had no knowledge of what was involved with that when I first started. Rest assured we're taking steps to fix this.

Quote from mikwuyma:

Ohhhh yeah, that's another thing. Searching prizes, challenges, and bidding wars by game name was incredibly awkward. Why isn't there an option to just type in the game name? You had to know part of the prize, challenge or bid name to efficiently find it.

Agreed, this is also something we're looking at streamlining. Also, I don't consider giving honest feedback 'shitting' mike Smiley

[quote="UraniumAnchor"]Java vs C# isn't going to make a big difference in the efficiency, the issue is that whatever database library we were using is either poorly optimized for the sort of traffic we were throwing at it, or we were using it incorrectly. Switching how the program interacts with the website will be a big help, so that shouldn't be an issue any more.[/quote]

Indeed.  The bottom line is that the program wasn't even designed to interact with a remote database in the first place, and I really had no knowledge of what was involved with that when I first started.  Rest assured we're taking steps to fix this.

[quote="mikwuyma"]Ohhhh yeah, that's another thing. Searching prizes, challenges, and bidding wars by game name was incredibly awkward. Why isn't there an option to just type in the game name? You had to know part of the prize, challenge or bid name to efficiently find it.[/quote]
Agreed, this is also something we're looking at streamlining.  Also, I don't consider giving honest feedback 'shitting' mike :)

dballin

2012-06-10 09:34:10 am

Balls jerky

Quote from mikwuyma:

Ohhhh yeah, that's another thing. Searching prizes, challenges, and bidding wars by game name was incredibly awkward. Why isn't there an option to just type in the game name? You had to know part of the prize, challenge or bid name to efficiently find it.

ditto x1000!! you should have tried it when there were 15 things named simply 'character choice.' ugggggghhhhhhhhh

dunnius

Edit history:

dunnius: 2012-06-10 12:01:51 pm

dunnius: 2012-06-10 12:01:36 pm

2012-06-10 11:48:25 am

Intruding N313 and F014

I may be able help with the back end (the database storage part) for this. I do database programming (mostly in C++) for a living. I could write server-side functions to help organize and gather the data, which I believe would help with the efficiency.

UraniumAnchor

2012-06-10 01:31:27 pm

Not a walrus

Nah, that's the one part that's pretty much on lock. Other than maybe adding a couple new columns. I appreciate the offer, though.

dunnius

2012-06-10 01:58:14 pm

Intruding N313 and F014

But it sounds like the client is making lots of server requests (or inefficient requests when there is lots of data) to get the data, which is causing it to run slow.

Rakuen

2012-06-10 02:17:44 pm

Weegee Time

Or maybe the tables are getting locked? I guess you could call that inefficiency...

SMK

2012-06-10 02:31:55 pm

Quote from dunnius:

But it sounds like the client is making lots of server requests (or inefficient requests when there is lots of data) to get the data, which is causing it to run slow.

The main problem is that the client is basically pulling in half the db for most requests (plus its round-tripping a lot for selecting sub-entities), because I wrote the client db code very lazily.