Project Idea, Input needed.

[Kreij]

What happens if after using the app a lot for a couple of years you end up with an XML file that is several hundereds of Megabytes (or a GB) in size. Would be kinda slow having to parse through the file to get what you need, no?

I've done some XML work, but not anything with really large files so I may be wrong.

@ Kreij↓ Ah, good. The way you wrote your other post had me wondering if it locked on parallel reads (which would definitely cause problems). My vote definitely goes for SQLite then, assuming this is a votable decision . No supplementary installs; just a single standalone executable that generates a single file database; definitely sounds nice to me...

I'm good with that. We don't need to vote, as we can make use of any or all DBs we want.
Sorry for the vague post, my bad.

As for a name I was kicking around "TPU Overload!"
(Unless W1zz bars me from using TPU or TechPowerUp! in the app. I don't want to have to deal with his lawyers. :/ )

Also, it's 0°F here at the moment, so I ain't going anywhere for awhile. lol

 

[xbonez]

What happens if after using the app a lot for a couple of years you end up with an XML file that is several hundereds of Megabytes (or a GB) in size. Would be kinda slow having to parse through the file to get what you need, no?

I've done some XML work, but not anything with really large files so I may be wrong.

Neither have I worked with XML files that huge, but the issue of the storage bloating up over years is something we'd face with any format, I believe.

 

[Kreij]

Databases directly access their tables through the SQL commands (and ignore the other data).
Can you do that with XML files?
Can you update portions of an XML file without re-writing the entire thing?

Like I said, I haven't done a whole lot with XML other that using to to store a little configuration data or read-only menu options.

Side Note : Do most of you have OpenOffice or Word? For reading requirement documentation, version update notes, etc. Hate to use Notepad. lol

 

[streetfighter 2]

I dunno much about parsing XML myself. Seems like it has lots of APIs and shouldn't be hard to implement-- if slow...?

EDIT: After doing some research I found an MSDN article about parsing large XML files that's faster than using DOM based APIs:
http://msdn.microsoft.com/en-us/library/ms950778.aspx
Unfortunately it's from 2004 and something better may exist.

I'm good with that. We don't need to vote, as we can make use of any or all DBs we want.

So you're thinking support for multiple databases? I think that's definitely the best option. It brings up more questions though, which would have come up anyway, about how we want to store user settings related to the GUI. If I wanted to load my database from SQLite and you wanted to get yours from MySQL, it would be pretty annoying to have to select the db source every time the program is started. The question then becomes: ini (files+1), registry, or some other option?

As for a name I was kicking around "TPU Overload!"
(Unless W1zz bars me from using TPU or TechPowerUp! in the app. I don't want to have to deal with his lawyers. :/ )

I like it. The logo literally draws itself.
Some other ideas: "TPUniverse", "TPUltra", "TransmitPU", "TransactPU". I like the last one because it actually makes sense
How about, "TPU Overlord"? It's a one letter change but it's awesome .

Do most of you have OpenOffice or Word? For reading requirement documentation, version update notes, etc. Hate to use Notepad. lol

No, I still cut all of my messages into stone tablets. My brother Josephus is claiming that he has invented something called "papyrus", the poor loon . . .

 

[xbonez]

Yeah, I've got to agree with you that XMl is more suited for config files and smaller files. Coming to think of it, it might not be the best idea for storage. Possibly, if you split up the data to be stored in multiple XML files, it would be a little more feasible, but having everything in a single file would be a waste of resources because deserializing the xml (AFAIK) cannot be done on parts of the file, or certainly can't be done as easily as deserializing the entire file.

 

[mlee49]

I want an Android TPU app. Naaoooooooww!

Anyone with app writing experience here?

 

[FordGT90Concept]

As long as we're talking about it, why not a segmented binary file? It's effectively what SQL does but eliminates the middleman (SQL). The only problem is that it can require a lot of memory to load the entire file into the memory. At the same time, the entire file often isn't required.

I prefer INI to XML in most cases because XML is horribly inefficient (tons of white space/wasted characters). The only times I wanted and did use XML is where users will be editing the file directly and there's information in there that isn't suitable for XML (like many objects under one type).

 

[Kreij]

I dunno much about parsing XML myself. Seems like it has lots of APIs and shouldn't be hard to implement-- if slow...?


So you're thinking support for multiple databases? I think that's definitely the best option. It brings up more questions though, which would have come up anyway, about how we want to store user settings related to the GUI. If I wanted to load my database from SQLite and you wanted to get yours from MySQL, it would be pretty annoying to have to select the db source every time the program is started. The question then becomes: ini (files+1), registry, or some other option?

My thought was to put all of the presisted GUI info into the datastore itself, and create an external file only containing the information to the datastore.
INI or XML is a non-issue. I agree with Ford, XML has a lot of overhead.

I like it. The logo literally draws itself.
How about, "TPU Overlord"? It's a one letter change but it's awesome .

I changed the name to that in the ReqDoc about 10 seconds before you posted this.
I guess we're on the same page there.
TPU Overlord - All your base are belong to us.

No, I still cut all of my messages into stone tablets. My brother Josephus is claiming that he has invented something called "papyrus", the poor loon . . .

No problem, OpenOffice just added "Export to Stone Tablet" in the latest version.

Yeah, I've got to agree with you that XMl is more suited for config files and smaller files. Coming to think of it, it might not be the best idea for storage. Possibly, if you split up the data to be stored in multiple XML files, it would be a little more feasible, but having everything in a single file would be a waste of resources because deserializing the xml (AFAIK) cannot be done on parts of the file, or certainly can't be done as easily as deserializing the entire file.

That's what I was worried about, but I am not an XML guru. Maybe more people can comment on this.

I want an Android TPU app. Naaoooooooww!
Anyone with app writing experience here?

No Android App for you until you pick a custom title. lawlz

 

[streetfighter 2]

My thought was to put all of the presisted GUI info into the datastore itself, and create an external file only containing the information to the datastore.
INI or XML is a non-issue. I agree with Ford, XML has a lot of overhead.

That sounds good to me.

Now how about skin support? Anyone know of free skinning tools for .NET?
I think I found one: http://www.componentfactory.com/

No problem, OpenOffice just added "Export to Stone Tablet" in the latest version.

Unfortunately that Export function requires the Quarry extension which I couldn't install because my computer has a broken slave controller.

 

[caleb]

JSON perhaps ? Works very well in files being constantly read over net. I actually switched from database's to flat files in JSON organized in day/week/month etc for my browser based reports application. I just do the sorting filtering with php built in array stuff. Report are a lot of columns tho and database engines dont like too many of these on the long run.

Check out tranmission torrent client. I have it on my router box and I think it uses JSON formatted config files.

 

[Kreij]

We'll worry about skins later.

I'm working on the requirement doc.
What functions do we want to include (choose what you like from OP or add something you want)?

We really don't need formal documentation, but I want some of the students here on TPU to get used to having to magically interpret what the end-users want from half-baked documentation made by people who have no clue of the actual process implementation.
(Notice : These ReqDocs will be more complete than you usually will see, and you can't get fired in the event you totally screw up. )
Welcome to real life coding.

As I stated in the OP, I want this to help those coming up in the field, not just another project for us long time code junkies.
If you already have a module and someone wants to take a shot at it, let them, and then mentor them. You may be surprised at the results.
We may actually release something, or we may not ... but people are going to learn along the way if they stick with it (myself included).

 

[streetfighter 2]

I know this kind of directly contradicts the "true programming experience" ethos but I think it will make discerning the project goals a little easier.

This is what I did on http://www.bubbl.us in a couple minutes. You can create a user account in a couple of seconds with a fake email and it's free. Add me as a contact, "streetfighter 2", or PM me and I'll add you to the edit list for the mind map.

Add any features you want and put your name as a dev for any features you want to be involved in implementing. Any features that end up without devs will be given a lower priority and ultimately might not be implemented.

BTW I forgot to add GUI. . .

If anyone disagrees with using bubbl, then disregard this.

 

[caleb]

I can try to model the data for databases and create DDLs once there are Req files.

Also I'd like some noob tasks to do in .NET so I can get into it.

 

[Kreij]

Looks good, we can try anything. Add Kreij please.
I will still post requirement docs as I want people to see what they may be up against.

 

[streetfighter 2]

I added you Kreij.

Also I'd like some noob tasks to do in .NET so I can get into it.

All tasks in .NET are noob tasks

Just pick any part that interests you and I'm sure there'll be people around to help you and pickup the slack.

 

[xbonez]

All tasks in .NET are noob tasks

Couldn't agree anymore.

Also, I like the functionality of keeping a track of all your attachments / screenshots uploaded.

 

[Nailezs]

subbed. i think this will be interesting

 

[Kreij]

Dang. If bubbl had a little more functionality it would perfect as an object oriented design tool.
But hey, it's free and does a decent job at flow visualization.

@xbonez : Software Repository Manager? <-- fancy title
Sounds good. Most of the coders will probably keep running revisions so they can fall back if needed, but it's nice to have someone who keeps it all.

I was contemplating assigning roles to people, but screw that. We're all now officially project managers, coders and artists.
When the first draft of the ReqDocs comes out, just pick what you want to work on.
I intend to keep this fun and stress free. We have enough things in life to deal with already.

The first thing we need is a list of the modules we want to include in the first revision.

 

[streetfighter 2]

Dang. If bubbl had a little more functionality it would perfect as an object oriented design tool.
But hey, it's free and does a decent job at flow visualization.

I noticed you made a little flowchart for the key manager. I was working on a flowchart of my own for the score thread manager, but I did it in another sheet. I'll add you to the permissions when next I logon.

This is also my first time using bubbl and the only feature that seems to be missing is the ability to switch between boxes (activities) and diamonds (decisions). Definitely useful anyway.

 

[Kreij]

I'm downloading the Community (Free) Edition of a tool called Visual Paradigm.
Looks cool and something new to play with.
(as if I don't have enought to do. DOH!)

 

[xbonez]

@Kreji: Haha, I like the title, but I didn't mean I would keep all revisions of the code (its best we do that in some online code repository). I meant, I'd like the tool we're building to have the option of keeping a track of all attachments you upload as a user.

 

[Kreij]

Oh. LOL ... That's the Upload Manager Module as stated in the OP.

 

[olithereal]

I was contemplating assigning roles to people, but screw that. We're all now officially project managers, coders and artists.
When the first draft of the ReqDocs comes out, just pick what you want to work on.
I intend to keep this fun and stress free. We have enough things in life to deal with already.

The first thing we need is a list of the modules we want to include in the first revision.

You might want to assign stuff to students, or just specify something like "**beginner/intermediate/experienced" on the roles description so that we don't get lost too easily.
my 2c.

 

[Kreij]

You may be right, olith, but I don't want to hinder someone from trying something even if it's over their head. You learn by getting out of your comfort zone.

Anyway .. what modules should we include? Which ones are useful? New ideas?
(You don't have to be a coder to reply to this)

 

[Nailezs]

Anyway .. what modules should we include? Which ones are useful? New ideas?
(You don't have to be a coder to reply to this)

can you elaborate? i dont understand what you mean.