Any thoughts on an online database?

[Pixel_Outlaw]

I was just thinking that maybe an online database could be created in order to help some of us fill out collections. Getting my AT&T board made me realize that there might be more out there than what meets the eye. I we were to all have access to a database it would be possible to search for models by number, mechanism, and branding. This might be a great centralized way to catalog the data between all the members here.


Thougts for data fields

Brand:
Model:
Mechanism:
Connector Plug Type:
Removable Cord Y/N:
Key Layout:
Additional info: (A short sentence or paragraph of outstanding features.)


I currently don't have enough experience to construct an online database but if we used accounts the same as we have here members could have a centralized hub of information. Just tossing out an idea here. I think that the additional info field would be especially useful.

[kishy]

I'd like to see such a thing happen but it'd be a nightmare I think.

Could also interconnect with or double-function as my suggested idea of a current resale value "blue book"

[wellington1869]

participation would be spotty probably.  On the other hand, if there were a way to "search sigs"?

but also i'm not sure what such a database would be for? just to catalog the types of boards out there?

[kishy]

That's great if you want to read through it, but meaning no offense whatsoever to the contributors, a proper database with nice front-end in which you can enter desired price range, desired switch, desired plug type etc. and get back a list of matching products is probably best.

It's not even that hard to do. I just don't know the actual SQL to do it, but I know it's not hard to do.

[ch_123]

I would like to see a proper wiki, where people can search brand names or part numbers or whatever... Obviously this would require work to implement and maintain...

[wellington1869]

That's great if you want to read through it, but meaning no offense whatsoever to the contributors, a proper database with nice front-end in which you can enter desired price range, desired switch, desired plug type etc. and get back a list of matching products is probably best.

It's not even that hard to do. I just don't know the actual SQL to do it, but I know it's not hard to do.

but what would be the purpose of the DB? buying and selling? (we already have a marketplace). Just a catalog? (would it be used much? the wiki is there already).  Requests to try out boards? (whats-his-name already created a website for that).

[Mental Hobbit]

but what would be the purpose of the DB?

Rid us of half page sigs?

[wellington1869]

Rid us of half page sigs?

what fun would that be? pimping out sigs is a major industry on GH
Sigs are also highly visible running advertisements in a way that a DB couldnt replace.

[wellington1869]

Is this like a DB of what people own, or a DB of historic and current boards available from manufacturers, retailers?  The wiki is supposed to be the latter, already.

If the DB has a really easy front end UI maybe people will eventually list their inventory with all the details - but also people buy/sell boards like crazy, would it be realistically kept up to date?

[kishy]

The DB as I'd envision it would be every keyboard ever known to any geekhack member. It would not be used to show off what we own but rather to document them. Same type of deal as current keyboard reference wiki but would allow easy cross-referencing by switch type, plug type, original computer it was shipped with, price range and so forth.

It would have a proper multiple table layout, not like some excel spreadsheet. Tables for manufacturers, connector types, every attribute of a keyboard. This allows for a maximum cross-referenceability...if that's a word.

Historical pricing reference would be a nice touch: known highest, known lowest, most recent and original would be a solid pricing guide. NIB could be excluded since NIB will generally (or should, at least) go as auction anyway.

This is something that requires thought and pre-planning. A proper database will have its full structure implemented in the beginning...I'd say that if less than a month of organized planning and thinking about structure and required fields went into it, it doesn't have sufficient pre-planning.

It would have a full-featured front end allowing lookup of all keyboards matching any particular attribute. Photos would be nice.

A technology similar to that app for Android that can do a search for something based on a provided photo rather than text would be awesome, but difficult as hell to implement I suspect. It'd need to compare a user provided photo against all photos of keyboards in the database then return records for matching units (would be useful for those no-name keyboards that don't even have FCC IDs)

[wellington1869]

The DB as I'd envision it would be every keyboard ever known to any geekhack member. It would not be used to show off what we own but rather to document them. Same type of deal as current keyboard reference wiki but would allow easy cross-referencing by switch type, plug type, original computer it was shipped with, price range and so forth.

i see. and i guess as a benefit, for instance, a person COULD use it to inventory what they own, by putting their name next to the db entry for that board.

so basically this is a proposal to replace the wiki's database-like info with a real database of every keyb we know about. maybe also cross-reference GH reviews from review section to the board. (which urls would work until the next vbulletin update anyway)

I can see the promise of it, but creation/maintenance is usu where large scale projects like this tend to break down. The nice thing about the wiki right now is there's little to break or maintain even if its largely incomplete for now.

that said, if some talented folks want to put in the time (even if just as an exercise), no harm in the experiment, right?  We wouldnt hold our breath until the finished product goes live, but its a nice idea for sure.

[kishy]

No, no member names. There could never be two entries for the same keyboard, it would completely destroy the idea of what this is.

If it's designed properly, zero maintenance needed. Provide a front end for easy keyboard entry and it maintains itself.

[wellington1869]

No, no member names. There could never be two entries for the same keyboard, it would completely destroy the idea of what this is.

well, i didnt mean actual names. But it could be a way for people to see who else owns it (or wants to try it, or is willing to let others try it, or is selling it, etc).
ie, so long as there is a database, why not put it to multiple uses, and eventually fold in the 'keyboard owner locator' and etc?

[kishy]

well, i didnt mean actual names. But it could be a way for people to see who else owns it (or wants to try it, or is willing to let others try it, or is selling it, etc).
ie, so long as there is a database, why not put it to multiple uses, and eventually fold in the 'keyboard owner locator' and etc?

That would work. Perhaps a field for users who own it, or perhaps a table of usernames corresponding to keyboard IDs that are owned (each keyboard would have a unique ID field in the db, allowing it to be referenced easily elsewhere)

[kishy]

Keyboard reference doesn't let you cross-reference by switch type and so forth. It presumes that you know the keyboard you want...what if you only know two or three details of the keyboard you want and are seeking to find a list of keyboards meeting those details?

[wellington1869]

Then you go to the most awesome Wiki entry EVER!

Link.

Addesso Blue Cherries now shipping for $70.  See the last post in the link.

excellent. any review on that yet? need review...

[kishy]

The current way we do things around here is about as counter-intuitive as Windows. Wiki articles are good for reading, they aren't good for looking stuff up.

[Pixel_Outlaw]

I've played with SQL a bit in college and honestly it makes stowing away data much easier. A wall of text is not an elegant solution for storing data.


My vision would be a database of hundreds of boards that ANY member can add to at any time. Get a new board? Just add it in.

If you saw something online and wanted more information because it was listed as "clicky" you could find the model in the database to verify that. The note area allows for owners to expand upon the essential description. databases have pretty much taken over spreadsheets because they hide the crap and only show what you need. The multi user addition ability is also great because there is no waiting for a single admin to update a page from a PM. This makes collecting data much more efficient. Even with ctrl+f method you are still going to have to slog through many results.

[wellington1869]

The current way we do things around here is about as counter-intuitive as Windows.

[Pixel_Outlaw]

I took an entry level course in college. I could make a very basic database but we never did learn how to incorporate online elements since the databases mostly were one user models that stayed on each student's computer.


Really it comes down to who has the skills to pay the bills. Parsing the sentences into the database would not be a huge problem with some basic string handling functions but again it might take somebody with more experience. If nobody here has the skills then it is not possible without some study. I was hoping we had a database programmer in our midst. I'm just a hobbyist game programmer at most. As a programmer, my mind is always thinking "optimize organize" ect. The database just seems like a natural fit for this because you have a series of similar entities (keyboards) that share the same attributes with unique configurations (field data).

[kishy]

I took an entry level course in college. I could make a very basic database but we never did learn how to incorporate online elements since the databases mostly were one user models that stayed on each student's computer.


Really it comes down to who has the skills to pay the bills. Parsing the sentences into the database would not be a huge problem with some basic string handling functions but again it might take somebody with more experience. If nobody here has the skills then it is not possible without some study. I was hoping we had a database programmer in our midst. I'm just a hobbyist game programmer at most. As a programmer, my mind is always thinking "optimize organize" ect. The database just seems like a natural fit for this because you have a series of similar entities (keyboards) that share the same attributes with unique configurations (field data).

I actually believe I could do it; this past semester I had an introductory PHP class and half the assignments ended up using PHP to interact with databases. It's my database integrity understanding that would hold me back, if anything, because I don't want to make a database of poor integrity and have issues pop up later.

[wellington1869]

this past semester I had an introductory PHP class

i need to take one of those. I keep putting it off. I was a Fortran wiz in college. Yea, I know, what a waste of tuition

[Eclairz]

Decided to create tables, its not complete and probably not efficient i'm sure you guys could improve on it. Hopefully this'll give you an idea on how to structure the database



Manufacturer
ID
Name
Year From
Year To


Keyboard Brand
ID
Name
Manufacturer
Manufacture Description
Year From
Year To


Keyboard
ID
Keyboard Brand
Part_Num
Weight
Dimensions
Interface
Detachable Cable
Year of Manufacture
Split or Whole
Keyboard Type (e.g. tenkeyless, standard)


Keyboard Switch Types
ID
Keyboard ID
Switch Type

Keyboard Colour
ID
Keyboard ID
Colour Primary
Colour Secondary

Keyboard Layout
ID
Keyboard ID
Layout

Keyboard Notes
ID
Keyboard ID
Notes
Created By
Created On

Keyboard Reviews
ID
Keyboard ID
Review Link
Created By
Created On

Keyboard Sold In
ID
Keyboard ID
Country

Keyboard Sold By
ID
Keyboard ID
Company Name
Company Link

Keyboard USB
ID
Keyboard ID
USB Type
Number Of
USB Type

Keyboard Misc Features
ID
Keyboard ID
Feature Type (e.g. Dip switches, pointing stick, trackball, etc)
Feature Name
Feature Desc

Keyboard Macro Keys
ID
Keyboard ID
Macro Key Type (e.g. FN+, dedicated key, programmable key, dip switch)

Keyboard Modifications
ID
Keyboard ID
Mod Name
Mod Desc
Mod Link

[lmnop]

I think this is a bad idea on our egos. ripster wouldn't have any threads to post in but who knows if hes even at the computer half the time maybe he has a device to make his presence known by posting random pictures

[Mental Hobbit]

No owner?

[ch_123]

Show Image

In *nix systems, there's a "yes" command, which continuously pipes 'y' into a specified program... I can only imagine that it was devised by some Unix hacker after seeing that episode of the Simpsons.

[kishy]

Well, that got my head spinning in circles.

Mental Hobbit, owner field would be bad, unless it were a field containing a list of every person who does own it. The idea is documenting the existence of the keyboard, not documenting who owns them.

I'm afraid of what will happen when people realize a 1391401 from 1987 is not the same as a 1391401 from 1993, thus creating the need for two 1391401s...

[ch_123]

It's a good thing they don't teach symbolic logic anymore in college. Gives me the edge.

They do in my college.

[ch_123]

Well, that got my head spinning in circles.

Mental Hobbit, owner field would be bad, unless it were a field containing a list of every person who does own it. The idea is documenting the existence of the keyboard, not documenting who owns them.

Yeah, who owns them is largely uninteresting. And I'm still think some sort of wiki would a be a good way of providing pictures, data, and a means of searching stuff when the user isn't 100% sure what they want.

[Mental Hobbit]

The idea is documenting the existence of the keyboard, not documenting who owns them.

I don't think we've agreed on that part yet. I'd like to have owners included.

[kishy]

I don't think we've agreed on that part yet. I'd like to have owners included.

What good does it do to identify who owns keyboards?

The only use for this sort of database that could justify spending any time at all to create such a database would be a tool for looking up any keyboard that ever existed according to some criteria. Ideally, once a keyboard was entered, the record would not be modified again assuming the info entered was correct.

[ch_123]

Ownership can be calculated by a simple algorithm anyway -

Code: [Select]

IF Manufacturer = "IBM"
  THEN Owner := "webwit"
  ELSE Owner := "Who cares"
END IF

[ch_123]

Technically my algorithm is wrong because it fails to consider the Industrial Model F...

[kishy]

Technically my algorithm is wrong because it fails to consider the Industrial Model F...

Webwit already owns it (as with all keyboards of the world). It's just a matter of it finding its way into his loving arms.

[itlnstln]

Too complicated.

[Mental Hobbit]

What good does it do to identify who owns keyboards?

You could contact members with questions on particular keyboards, especially rare ones.
Besides, getting an overwiew of all my crap would be a major incentive to enter my boards into the database.

I'm getting too many PMs already.  Leave the owner part out.  That's what sigs are for.

Leave the field empty then. And turn off PMs in the userCP.