|
Dakota_Fanning
Hungary2349 Posts
For those who are interested here are some info about my new project that I've been working on in the past 5 months.
The new project is the successor of Sc2gears™.
(I let its name still be a secret for now.)
I started the successor from ground zero almost 5 months ago.
But wait...
What has led to a successor instead of just updating Sc2gears™?
Many things. To mention the main reasons:
- Sc2gears has become big. Too big. I never thought I will add so many features to it in the beginning and so it was becoming harder and harder to maintain.
- Many claimed Sc2gears was difficult to use. It has so many features, many didn't understand a lot of them.
- I have also received complaints about the UI design of Sc2gears.
- Blizzard finally decided to provide official support for parsing replays: they released a low-level open source python library called 's2protocol'.
- During the last 3-year development and maintenance of Sc2gears, I improved a lot and gained an enormous experience, I got to know the SC2 replay format more and more, and Blizzard included more and more information in replays.
- Java 7 became standard, and it has new features which can be put to use.
Due to the aforementioned improvements, if I were to rewrite Sc2gears, I would do many things quite differently. It is much easier to do these different things in a fresh new project than to refactor an existing huge and complex project. Without refactoring on the long-run I would just make my job even worse.
No question it would've been easier to just update Sc2gears than to create something entirely new. But I plan a much longer and more important role for my project than to just choose the easier path.
Main design aspects of the successor of Sc2gears:
- RISK: Rethink everything (from Sc2gears). Invent new things. Simplify the complicated things. Keep the good things.
- Unified and more user friendly interface. Be intuitive, productive, and provide context sensitive help and tips for those who need it.
- Users have quite different computer skills. Make the user interface, the displayed information, options, settings and controls scalable based on this.
- Even more customization and configuration than in Sc2gears, but also with a more pleasing presentation; it should be easier to find the settings you're looking for.
- Provide the deepest and most comprehensive insight into replays both on low level and on high level.
- Be professional.
- Much higher quality overall. Something that is easy, fun and profitable to use.
I think this is all I wanted to say for now.
If you have questions, suggestions, I'm open to them (e.g. what are the things you most missed from Sc2gears, the things that you loved about Sc2gears, the things that most bothered you in Sc2gears etc.). Every suggestion helps making the successor better.
|
Can't wait to see and try it :D
|
I look forward to it. I'm actually planning on using SC2 gears to create a corpus for sc2 replays, so this will only help me even more. I love the tag cloud feature so much.
|
I'll be honest, I only use sc2 gears for APM alert. Just keep that in and I'll be happy ^_^
|
|
|
As a professional software engineer I can't help but wonder why you don't choose for a web based application, care to elaborate?
And if you are worried about the version of Java installed, why not bundle the application with a JRE? With all the security exploits lately (fucking Oracle is wrecking Java even worse than Sun could), many people have been advised to uninstall Java anyway.
Good luck, looking forward to a new version 
|
Dakota_Fanning
Hungary2349 Posts
On July 30 2013 22:09 TaphrWesley wrote:As a professional software engineer I can't help but wonder why you don't choose for a web based application, care to elaborate? And if you are worried about the version of Java installed, why not bundle the application with a JRE? With all the security exploits lately (fucking Oracle is wrecking Java even worse than Sun could), many people have been advised to uninstall Java anyway. Good luck, looking forward to a new version
To mention a few reasons that immediately pops into my mind:
-A web based application can't analyze thousands of replays located on your computer instantly.
-A web page can't even access them.
-Many users don't want to upload their replays to a server.
-Even those who would agree to upload their replays would not want to pay for the resources it requires to store and/or process them.
-A web based application can't provide you services like replay auto-save, APM alert, On-top info dialogs etc.
-A web based application is slower than a desktop app.
Don't get me wrong, I'm not against web programming (doing that for a living for 10 years), but there are things which are better off the web.
|
I love the MMR stats plugin so I hope you're in contact with those guys to get support for it into the new tool as well.
|
Just spit-balling ideas since you said you were open to suggestions.
What if it a user could create a build order "template" (according to some predetermined optimized timings) and compare the template to data pulled from relevant replays (tag the replays some how?). It would be interesting to visually see improvement in game play "accuracy" insofar as performing basic build orders are concerned.
Taking it a step further, what if advanced players could then look at their replays and using all the other tools given in the program, assess why they deviated from the build order with respect to specific events in-game. Using this information these players can go back and add a modification or even a "branch" to the template (and perhaps be able to note why a deviation or change was needed). This could ultimately result in a build order "tree" with branches dictated by reactions to the opponents build/game play. This would then be useable as a template for future comparisons.
I'm not sure how best this can be visualized other than a basic tree diagram going from left to right with the horizontal axis being game time and vertical axis used for branching space.
The neat part would be this would be useful for both players trying to master the most basic builds, all the way up to the most complex builds. Not mention it offers the possibility of being able to visualize and extend build orders beyond say the typical 10-15 minute or so mark (where begins to get hazy due to specific in game events) by helping with long-term pattern recognition.
I guess simply put, we normally deal with static build orders that are more or less independently designed relative to the opponents behavior or various other circumstances. I'm suggesting being able to map the dynamic build order with consideration given to the "if" factor. If you scout this, if you sustain x amount of damage here, if you expand here, if you on this map with these spawns. Make a note of it on a branch and map it out. Player can organize different trees as they see fit. For example by map, race, spawn, etc.
I hope I was able to explain my idea clear enough. Also, my programming background is limited so if what I just suggested requires a 50 member team and million dollars with 4 year development time to complete, by all means ignore me. 
|
Sounds really interesting. I'm looking forward to it!
|
Finland7871 Posts
|
I dont think SC2 gears is difficult to use... there are many options and huge menus, it is overall too much, but that is just question of interface design. (Mouse cursor game, ingame apm alert are jsut over the top, beyond "nice to have")
My little idea: make it possible to put some stats on the start page like: APM over the last 100 games, winrate over the last 10 games, best map this season etc...
Also: Opponent database to allow easy searching and takeing notes (I hate being allined by the same guy  )
|
what is bwhf btw.
i am curious...
i have been a fan of sc2gears since first touch and i guess finally one day your programme will be part of starcraft 2
|
You suck at keeping your project secret. And we love you for it <3
|
Dakota_Fanning
Hungary2349 Posts
On July 30 2013 22:55 winthrop wrote:
what is bwhf btw.
i am curious...
i have been a fan of sc2gears since first touch and i guess finally one day your programme will be part of starcraft 2
First hit for Googleing BWHF:
https://code.google.com/p/bwhf/
StarCraft BroodWar Hacker Finder, anti-hack, replay analyzer-organizer and utility tool
It was a hacker finder and replay analyzer-organizer tool for StarCraft BroodWar with a central hacker database and a Player's Network. Open source (including the server side!) with a 60 page developer documentation.
|
On July 30 2013 23:04 Dakota_Fanning wrote:Show nested quote +On July 30 2013 22:55 winthrop wrote:
what is bwhf btw.
i am curious...
i have been a fan of sc2gears since first touch and i guess finally one day your programme will be part of starcraft 2 First hit for Googleing BWHF: https://code.google.com/p/bwhf/StarCraft BroodWar Hacker Finder, anti-hack, replay analyzer-organizer and utility tool It was a hacker finder and replay analyzer-organizer tool for StarCraft BroodWar with a central hacker database and a Player's Network. Open source (including the server side!) with a 60 page developer documentation.
Did you ever think about an antihack tool for SC2? Is it even possible?
|
Damn, keeping us a bit in the dark there. Always loved using SC2Gears, so really excited to see what's next! :D
|
Dakota_Fanning
Hungary2349 Posts
On July 30 2013 23:06 graNite wrote:Show nested quote +On July 30 2013 23:04 Dakota_Fanning wrote:On July 30 2013 22:55 winthrop wrote:
what is bwhf btw.
i am curious...
i have been a fan of sc2gears since first touch and i guess finally one day your programme will be part of starcraft 2 First hit for Googleing BWHF: https://code.google.com/p/bwhf/StarCraft BroodWar Hacker Finder, anti-hack, replay analyzer-organizer and utility tool It was a hacker finder and replay analyzer-organizer tool for StarCraft BroodWar with a central hacker database and a Player's Network. Open source (including the server side!) with a 60 page developer documentation. Did you ever thought about an antihack tool for SC2? Is it even possible?
Yes, I have. To an extent, yes. But I think the time is not (yet) right.
|
|
|
|
|
|
|
|
|
|
EPT
