Forms 'n Molds

[shmax]

Another development over the last month or so has been a refinement of the "forms" system. Toys can now have up to 3 different forms assigned to them. So a triple-changer like Astrotrain is tagged as being 1. a robot, 2. a space shuttle, and 3. a train. When browsing by form, he'll turn up in all 3 of those categories. What does this mean to you, the power-user? Well, it makes identification of toys and parts a bit easier, because now if you've got a toy in your lap and you're not sure if it's a car or some kind of two-headed squirrel, you can try either one and chances are good it'll turn up. I hope you guys will get some use out of this feature, but be warned that the first person who chimes in with "hey! what about those 10-changer toys?" will be pelted with Go-Bots.

The observant among  you will also have noticed that the form information is now separate from the mold information. The mold system has, in fact, been totally rebuilt. With the new system, you can see the origin and complete history of any part by carefully examining its mold path. For example, take a look at this familiar face:

http://www.shmax.com/part_details/432

See over on the right, where it says "Mold:"? Note that his path begins with "Diaclone", which is the grand-papa of the original G1 Transformers toyline. So with this system, each toy's mold path will contain the name of the toyline where it first appeared. Want to know if something is a repaint or reissue of something else? Just examine its mold path.

Think that's nerdy? We're just getting started! You can even see if a toy is a retool (or a retool of a retool) of another toy by examining its mold path. Parts that are retools will have the word "retool" in their mold path.

Going even further, you can browse the entire database by mold, by clicking on the "mold" mini-tab over on the left. An interesting feature that emerged is that if you want to see only the molds that originated in a particular toyline, you only need to browse to that toyline name in the mold tree. For example, here are all the molds that were introduced in the new "Universe" deluxe line:

http://www.shmax.com/Database/4101?browse=2

You need never be fooled again by the repaint fairy!

[Chaotic Descent]

but be warned that the first person who chimes in with "hey! what about those 10-changer toys?" will be pelted with Go-Bots.

I'm not sure I understand why you went with a fixed maximum of 3 rather than a variable number that allows for all existing figures. Figures with more than 3 modes are not that small a minority. Energon Autobots alone have 4 modes, not counting the combinations with other figures. (top half and bottom half in a combination does count as two distinct modes) Same with the Energon 5-member gestalts. (boy, Energon was full of combiners with 2 different combiner modes!)
So where are my free Go-Bots?

Well, I picked up an AMP package again... (Apache, MySQL, PHP) I think it might be easier this time, for some reason. I'm under Linux (Ubuntu). I'd really like to design some nice ways to display the data, starting with the layout I had for my collection years and years ago. That'll be my first big goal. you know... if I actually get anywhere.
I know I've pretty much done nothing since I came here but... hey, that's me.

[shmax]

Well, I did it the way I did for a number of reasons:

1. Query time. For every form I add, the time it takes to run a form query on the figure/parts database increases exponentially.

2. UI considerations. I would have to toss out my current technique of assigning forms and come up with something entirely new, to support an arbitrary number of forms.

3. Schema complexity. Right now the three forms are simply columns in a parts table. For your method, I would have to create a new table, thus adding even more complexity in the form of additional joins, and I'm already struggling with query time and complexity as it is.

But nothing's set in stone. In short, I did it this way because it was the fastest, easiest, cleanest way to do it. I didn't think that the handful of known 5, 6, 7, 8, 9, and 10-changers really warranted the extra work and complexity. I really only added the 3rd form so that identification of triple-changers would be more robust (ie. searching for "train" or "STS" will get you Astrotrain). I just can't imagine someone trying to id a 10-changer by searching for one of its obscure kind-of-pushing-it-forms (ie. "tank-like thing with legs sticking out of it").

But I'll consider this a vote for going the extra mile. If more people chime in with similar opinions, I'll take the time to look into it.

Max

but be warned that the first person who chimes in with "hey! what about those 10-changer toys?" will be pelted with Go-Bots.

I'm not sure I understand why you went with a fixed maximum of 3 rather than a variable number that allows for all existing figures. Figures with more than 3 modes are not that small a minority. Energon Autobots alone have 4 modes, not counting the combinations with other figures. (top half and bottom half in a combination does count as two distinct modes) Same with the Energon 5-member gestalts. (boy, Energon was full of combiners with 2 different combiner modes!)
So where are my free Go-Bots?

Well, I picked up an AMP package again... (Apache, MySQL, PHP) I think it might be easier this time, for some reason. I'm under Linux (Ubuntu). I'd really like to design some nice ways to display the data, starting with the layout I had for my collection years and years ago. That'll be my first big goal. you know... if I actually get anywhere.
I know I've pretty much done nothing since I came here but... hey, that's me.

[Chaotic Descent]

Well, I did it the way I did for a number of reasons:

1. Query time. For every form I add, the time it takes to run a form query on the figure/parts database increases exponentially.

2. UI considerations. I would have to toss out my current technique of assigning forms and come up with something entirely new, to support an arbitrary number of forms.

3. Schema complexity. Right now the three forms are simply columns in a parts table. For your method, I would have to create a new table, thus adding even more complexity in the form of additional joins, and I'm already struggling with query time and complexity as it is.

But nothing's set in stone. In short, I did it this way because it was the fastest, easiest, cleanest way to do it. I didn't think that the handful of known 5, 6, 7, 8, 9, and 10-changers really warranted the extra work and complexity. I really only added the 3rd form so that identification of triple-changers would be more robust (ie. searching for "train" or "STS" will get you Astrotrain). I just can't imagine someone trying to id a 10-changer by searching for one of its obscure kind-of-pushing-it-forms (ie. "tank-like thing with legs sticking out of it").

But I'll consider this a vote for going the extra mile. If more people chime in with similar opinions, I'll take the time to look into it.

Max

Well, I was thinking there wouldn't be much use in it for combiners, but being able to search for combiner components might be useful after all.

1) doesn't it only look in certain columns for forms? In that case, if you had a separate table for forms, wouldn't a search for forms only search in a form table? then you wouldn't exponentially be searching for each new form. You also wouldn't be searching for third forms on all the TFs that only have two forms. Doesn't the savings in time balance out? What, do you run calculations on which is faster?
Also, couldn't you have pre-something-ed results saved for at least the most popular searches, and updated when the database is updated? I know TFU.info is totally pre-whatever-ed. or... actually, I think they're manually edited, since they often end up with errors. :S

I dunno... I'm always in favor of accuracy.

I was thinking about maybe different versions of the database, one simplified from the other, but the only thing I can think of that collectors wouldn't really need is tech specs and bios, and TFwiki info such as cartoon and comic appearances. ("Why do that when we have TFwiki? Just link to them." I guess...) Also obscure info like plastic types and color codes if it were possible to accurately figure those out.

[shmax]

1) doesn't it only look in certain columns for forms? In that case, if you had a separate table for forms, wouldn't a search for forms only search in a form table? then you wouldn't exponentially be searching for each new form. You also wouldn't be searching for third forms on all the TFs that only have two forms. Doesn't the savings in time balance out? What, do you run calculations on which is faster?
Also, couldn't you have pre-something-ed results saved for at least the most popular searches, and updated when the database is updated? I know TFU.info is totally pre-whatever-ed. or... actually, I think they're manually edited, since they often end up with errors. :S

Okay, you got me. It's not actually an exponential increase. I guess the performance gripes aren't the real issue--the main problem is that I would have to do some fancy UI work on the input forms. Or, I guess I could just have 10 form widgets on every page, but that might clutter things up a bit. Tell ya what, I'm in the middle of a few other things right now, but since you've been persistent I'll make a point of looking into this a bit further when I can. Good enough?

I was thinking about maybe different versions of the database, one simplified from the other, but the only thing I can think of that collectors wouldn't really need is tech specs and bios, and TFwiki info such as cartoon and comic appearances. ("Why do that when we have TFwiki? Just link to them." I guess...) Also obscure info like plastic types and color codes if it were possible to accurately figure those out.

If you're willing to become a moderator and personally research and input info like plastic types and color codes for around 20,000 parts and pieces, then I'll be only too glad to engineer those features. As for having different versions of the database, I'm not sure what you're after. Are you saying you feel like there's too much information?

[DarkConvoy]

The Mold thing is too confusing. Also the adding a figure is too difficult as it takes like 5 tries before I could get a new fig in.

[shmax]

Yes, the mold thing is admittedly a bit complex. It will get rewritten one of these days to be a little friendlier. In the meantime, I've started work on a FAQ for the whole site, where I'll try to explain things like the mold widget in more depth. Failing all that, you can always simply leave the mold widget alone and it will always be filled in by a moderator.

As for adding a figure, well, what exactly are you having trouble with? Aside from the mold widget, the rest should be fairly straightforward!

The Mold thing is too confusing. Also the adding a figure is too difficult as it takes like 5 tries before I could get a new fig in.