spacecombat-develop Mailing List for SpaceCombat
Status: Beta
Brought to you by:
kahless
Menu
▾
▴
spacecombat-develop
You can subscribe to this list here.
| 2003 |
Jan
|
Feb
|
Mar
|
Apr
|
May
(20) |
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2004 |
Jan
|
Feb
|
Mar
|
Apr
|
May
(1) |
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|
From: <ben...@id...> - 2004-05-22 13:49:55
|
Dear Open Source developer I am doing a research project on "Fun and Software Development" in which I kindly invite you to participate. You will find the online survey under http://fasd.ethz.ch/qsf/. The questionnaire consists of 53 questions and you will need about 15 minutes to complete it. With the FASD project (Fun and Software Development) we want to define the motivational significance of fun when software developers decide to engage in Open Source projects. What is special about our research project is that a similar survey is planned with software developers in commercial firms. This procedure allows the immediate comparison between the involved individuals and the conditions of production of these two development models. Thus we hope to obtain substantial new insights to the phenomenon of Open Source Development. With many thanks for your participation, Benno Luthiger PS: The results of the survey will be published under http://www.isu.unizh.ch/fuehrung/blprojects/FASD/. We have set up the mailing list fa...@we... for this study. Please see http://fasd.ethz.ch/qsf/mailinglist_en.html for registration to this mailing list. _______________________________________________________________________ Benno Luthiger Swiss Federal Institute of Technology Zurich 8092 Zurich Mail: benno.luthiger(at)id.ethz.ch _______________________________________________________________________ |
|
From: Edward N. <cry...@gm...> - 2003-05-28 23:55:23
|
Hello spacecombat-develop, How is the development going at the moment? Ive not heard much from the list lately. -- Best regards, Edward mailto:cry...@gm... |
|
From: Herbert P. <ka...@yo...> - 2003-05-19 12:50:21
|
Hi, I've just made a basic structure for SpaceCombat. I'm using Velocity and Postgresql. You can see the souce code at http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/spacecombat/SpaceCombat2/ (you should be able to check it out using your sourceforge username & password.. if i forgot someone to add to the developer list, please tell me.)A running version is at http://yourhell.com:8090/SpaceCombat2/SpaceCombat ... not really spectacular, because there is nothing yet .. Every main section should get one servlet (extending SpaceCombatServlet) (e.g. Ships, Planets, etc. i suggest also making a package for these things) I've also created a class called SpaceCombatSubServlet which can be used for e.g. Planets/Buildings or something. (You see an example at http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/spacecombat/SpaceCombat2/work/net/spacecombat/game/admin/) In the theme files you can use #string( "whatever" ) to get the string with the key 'whatever' from the database in the language the user has choosen.. (languages can be added with the admin interface) Please tell me what you think about that... cu Herbert |
|
From: Herbert P. <ka...@yo...> - 2003-05-15 15:09:29
|
Hi, because the starmap is (one of) the most important feature in SpaceCombat i would like to be sure that it is possible to make a good looking and more or less fast (it should also work on not high end notebooks :) ) in java which is really useable (not just for getting an idea of the universe, but it should also be used to navigate ships, and select planets, etc.) and because ships can navigated to each coordinate (not only planets) it may be necessarry to make something like in most 3d grafic programs .. view from top, left, front.. well .. sounds a little complicated ? .. maybe it is more fun to play if we use a 2d starmap again ? what do you think ? (btw. the starmap should not only display planets.. also scanrange of them and ships .. maybe also the weaponrange of the ships) cu Herbert |
|
From: James H. <ja...@is...> - 2003-05-14 16:43:53
|
On Wed, May 14, 2003 at 07:59:49PM +0200, Thomas Haller wrote: > I was thinking of the > ticker_starts_calculations-db_does_calculations-ticker_ends_calculations > response time. Calculations are (p.e.): builder has completed building > -> make new entry in db and start new building from buildingqueue. This > response time (over all calculations) has to be (for a slow > rt-strategy-game) less then 10 secs. In the current game it is 10mins. The way I've got around that is to extend java.util.TimerTask and use java.util.Timer's to kick off events. Timers are very lightweight and the documents say that you can use thousands of timers without a problem. Although this is getting into implementation stuff, it's worth pointing out that it's doable. -- When a true genius appears in the world, you may know him by this sign, that the dunces are all in confederacy against him. - Jonathan Swift |
|
From: Kahless <ka...@sp...> - 2003-05-14 16:09:36
|
On Wed, 14 May 2003 19:59:49 +0200 Thomas Haller <ba...@fe...> wrote: > Thats not exactly, what i was speaking of. I didnt think of the > client-db-client response time, where i agree with you, that it will be > fast enough. I was thinking of the > ticker_starts_calculations-db_does_calculations-ticker_ends_calculations > response time. Calculations are (p.e.): builder has completed building > -> make new entry in db and start new building from buildingqueue. This > response time (over all calculations) has to be (for a slow > rt-strategy-game) less then 10 secs. In the current game it is 10mins. > > The 10 minutes aren't because of the database, but because of the calculations made by the ticker ... it wouldn't be much faster if the data would be in RAM instead of using a database ... thats the reason why i would like to make it realtime .. without such a centralized ticker which has to do all calculations at once .. |
|
From: Thomas H. <ba...@fe...> - 2003-05-14 15:59:50
|
James Harlow wrote: >On Wed, May 14, 2003 at 01:01:41PM +0200, Thomas Haller wrote: > > >>Herbert wrote, that he (is / thinks about) designing a db. Are u still >>planning to make a realtime game?(= 1tick within secs) Then i think u >>cannt do both, using a db and make the ticker fast enough.I see the only >>way of a rt game in making the ticker a persistent process, that has all >>data in his memory. Or does someone have experience in rt applications >>using dbs? >> >> > >Yeah, you can definitely do both, I've done a 30,000 player db-driven game >that had a latency on the server of less than a tenth of a second. Any >half-way reasonable db wil cache most of the data in RAM and optimise >frequent queries away if it can, so the cost of using a db will be >almost as small as possible. > > Thats not exactly, what i was speaking of. I didnt think of the client-db-client response time, where i agree with you, that it will be fast enough. I was thinking of the ticker_starts_calculations-db_does_calculations-ticker_ends_calculations response time. Calculations are (p.e.): builder has completed building -> make new entry in db and start new building from buildingqueue. This response time (over all calculations) has to be (for a slow rt-strategy-game) less then 10 secs. In the current game it is 10mins. |
|
From: Kahless <ka...@sp...> - 2003-05-14 15:19:06
|
On Wed, 14 May 2003 15:54:34 +0100 James Harlow <ja...@is...> wrote: > > The main problem will be the gateway, no OS I know can do 30,000 > network connections efficiently (well, maybe VxWorks can, but I've never > used it). So what we have to do is write the game either with a push > mechanism so that the client is informed of interesting events (not all > of them) or write it so that players aren't seriously disadvantaged by > not updating as frequently as possible. It may turn out that these are > the same thing. Well, for me it is not so important that the player sees the current status in realtime, than that the calculations are in realtime and not once per hour. Those nice little javascripts which updates a countdown, when a construction is done should be enough ... i think it is not necessarry to have e.g. the 3d starmap updated in realtime, because actions with ships and planets are done pretty slow .. so it is a strategy game, not an action game .. and i think it should (also) be playable that you take a quick look at your race when you are bored at work/school to give some commands :) so the process of the game may be very slow .. flying from one planet to the closest neibour planet should take about 3 hours or something and fights between ships should also take a few hours until there is a winner/looser .. (of course.. it would also be nice if there is some sort of constant which can be changed to change the whole thing into an action game ;) ) cu Herbert |
|
From: Edward N. <cry...@gm...> - 2003-05-14 15:10:39
|
Hello James, Wednesday, May 14, 2003, 3:54:34 PM, you wrote: JH> On Wed, May 14, 2003 at 01:01:41PM +0200, Thomas Haller wrote: >> Herbert wrote, that he (is / thinks about) designing a db. Are u still >> planning to make a realtime game?(= 1tick within secs) Then i think u >> cannt do both, using a db and make the ticker fast enough.I see the only >> way of a rt game in making the ticker a persistent process, that has all >> data in his memory. Or does someone have experience in rt applications >> using dbs? JH> Yeah, you can definitely do both, I've done a 30,000 player db-driven game JH> that had a latency on the server of less than a tenth of a second. Any JH> half-way reasonable db wil cache most of the data in RAM and optimise JH> frequent queries away if it can, so the cost of using a db will be JH> almost as small as possible. JH> The main problem will be the gateway, no OS I know can do 30,000 JH> network connections efficiently (well, maybe VxWorks can, but I've never JH> used it). So what we have to do is write the game either with a push JH> mechanism so that the client is informed of interesting events (not all JH> of them) or write it so that players aren't seriously disadvantaged by JH> not updating as frequently as possible. It may turn out that these are JH> the same thing. why not have many gateways? split the db onto three computers organised by user name, if that is what we are storing, a-j, k-p, p-z. we can have a game server which tells the client which gateway to connect to. the connections could be udp, then we do not need persistance afaik.. -- Best regards, Edward mailto:cry...@gm... |
|
From: James H. <ja...@is...> - 2003-05-14 14:54:40
|
On Wed, May 14, 2003 at 01:01:41PM +0200, Thomas Haller wrote: > Herbert wrote, that he (is / thinks about) designing a db. Are u still > planning to make a realtime game?(= 1tick within secs) Then i think u > cannt do both, using a db and make the ticker fast enough.I see the only > way of a rt game in making the ticker a persistent process, that has all > data in his memory. Or does someone have experience in rt applications > using dbs? Yeah, you can definitely do both, I've done a 30,000 player db-driven game that had a latency on the server of less than a tenth of a second. Any half-way reasonable db wil cache most of the data in RAM and optimise frequent queries away if it can, so the cost of using a db will be almost as small as possible. The main problem will be the gateway, no OS I know can do 30,000 network connections efficiently (well, maybe VxWorks can, but I've never used it). So what we have to do is write the game either with a push mechanism so that the client is informed of interesting events (not all of them) or write it so that players aren't seriously disadvantaged by not updating as frequently as possible. It may turn out that these are the same thing. -- When a true genius appears in the world, you may know him by this sign, that the dunces are all in confederacy against him. - Jonathan Swift |
|
From: Thomas H. <ba...@fe...> - 2003-05-14 09:01:43
|
Herbert wrote, that he (is / thinks about) designing a db. Are u still planning to make a realtime game?(= 1tick within secs) Then i think u cannt do both, using a db and make the ticker fast enough.I see the only way of a rt game in making the ticker a persistent process, that has all data in his memory. Or does someone have experience in rt applications using dbs? |
|
From: Herbert P. <ka...@yo...> - 2003-05-14 07:21:01
|
Hi, > What life cycle are we using for this? > > How long do we expect this project to be in the design phase? Thats > what its in isnt it? Well.. yes.. it is in the design phase .. and it should be as short as possible .. I'm currently playing around with servlets to get some sort of basis, where other servlets can attached. But as I allready said I have no real expierence with servlets so it can take a few days ... However... there need to be more things to be clarified, also on the game design.Btw. we need to define some game features, so we can start making a database structure.I suggest using most of the features we wanted for SC final: - players & races: I think we should make the design flexible enough to allow players to have more than one race. Races have specified attributes (defines the strength in fightings, adaptability to planet conditions, speed of research and speed of construction of ships & buildings), and perfect planet conditions which specifies how fast their population grow on a planet. (oxygen content, gravitation and temperature) - variable amount of resources (resources are stored in their own table, not in the ships & planets table) - players can build their own ships .. so we need a shipclasses (Ships are split up into classes, like fighters, transporters, etc.), shipdesign (for the basic designs for ships, where the player can place equipment into the slots), shipdesignslots, equipment (equipment available to the players), shiptypes (designs with equipment in their slots..), shipequipment (equipment place into slots of a shiptype by a user), ships (build shiptypes) - research .. as Id for techtree items we used a blob where we simply made a binary id (e.g. first element is 1 (1), second is 2 (10), third is 4 (100), then 8 (1000), etc. .. so when something depends on the third and fourth item, it has the dependency 12 (1100) .. it isn't that nice in the database design, but easy and fast to use. - starmap .. i think we should try using a 3d starmap.. each planet and ship can only be at one coordinate, but on one coordinate can be infinite ships and planets at once. players should see all planets in the starmap, but only get informations about them when they are in scanrange .. ships should only be seen within scanrange... maybe we should split the starmap into different sections, so it is more variable on the player numbers.. (when one section is full, we simply create another one) .. the sectors could be by wormholes or something in the center of each section. (just an idea .. never used that) hmm.. i hope i didn't forget something important .. cu Herbert |
|
From: Edward N. <cry...@gm...> - 2003-05-13 23:58:10
|
Hello spacecombat-develop, What life cycle are we using for this? How long do we expect this project to be in the design phase? Thats what its in isnt it? -- Best regards, Edward mailto:cry...@gm... |
|
From: Kahless <ka...@sp...> - 2003-05-12 22:31:10
|
On Mon, 12 May 2003 21:53:36 +0100 James Harlow <ja...@is...> wrote: > > And does anyone know a good template library we could use ? i found > > two which looks nice: > > http://www.webmacro.org and > > http://jakarta.apache.org/velocity/ > > I've heard good things about velocity. > > > (or does anyone prefer jsp ? imho servlets have a nicer structure > > because i think code shouldn't be in an html document... as well as > > there should be no html in code) > > I think that jsp's that get their information from servlets is a good > idea, it's just like having a rendering client that lives on the server. > But because i would like to make SpaceCombat theme based i think it would be better to load html templates from in the code. When a user change his theme, the servlet simply loads the files from another directory. Btw. when making a servlet. what is the more common way in such a more or less big project. using one servlet which loads all the other classes (e.g. http://mydomain.com/SpaceCombat?section=ships;method=build), or making one servlet for every 'section' (like ships, planets, research, etc.- e.g. http://mydomain.com/SpaceCombat/Ships?method=build) In perl i'm used to make one index.pl which loads all the other modules, because i need to initialize all on every http request from new..(configuration, db connection, user configuration, etc.) cu Herbert |
|
From: James H. <ja...@is...> - 2003-05-12 20:53:41
|
On Mon, May 12, 2003 at 10:34:18PM +0200, Kahless wrote: > On Mon, 12 May 2003 18:43:54 +0100 > James Harlow <ja...@is...> wrote: > > > > > How should the interface for the servlet work between server and > > > client? > > > > I don't have any firm suggestions for the interface, but a lot of > > servlets use HTTP. > > I meant which protocol when we make an external 3d client in addition to a > browser based version. There's no reason to change from http - either to retrieve a file like latest.php?user=hythloday, or return a the data inside the http header. > And does anyone know a good template library we could use ? i found two > which looks nice: > http://www.webmacro.org and > http://jakarta.apache.org/velocity/ I've heard good things about velocity. > (or does anyone prefer jsp ? imho servlets have a nicer structure because > i think code shouldn't be in an html document... as well as there > should be no html in code) I think that jsp's that get their information from servlets is a good idea, it's just like having a rendering client that lives on the server. -- When a true genius appears in the world, you may know him by this sign, that the dunces are all in confederacy against him. - Jonathan Swift |
|
From: Kahless <ka...@sp...> - 2003-05-12 20:35:44
|
On Mon, 12 May 2003 18:43:54 +0100 James Harlow <ja...@is...> wrote: > > How should the interface for the servlet work between server and > > client? > > I don't have any firm suggestions for the interface, but a lot of > servlets use HTTP. I meant which protocol when we make an external 3d client in addition to a browser based version. And does anyone know a good template library we could use ? i found two which looks nice: http://www.webmacro.org and http://jakarta.apache.org/velocity/ (or does anyone prefer jsp ? imho servlets have a nicer structure because i think code shouldn't be in an html document... as well as there should be no html in code) cu Herbert P.S.: I just noticed, that i made a mistake in the configuration of the mailinglist .. mailman didn't set the Reply-To field .. so when replying to a message it was directed to the sender instead of the list .. i hope it works now .. sorry for that. |
|
From: Senthil k. <s_s...@si...> - 2003-05-12 09:35:27
|
Hi, Does 3 tier architecture mean use of Enterprise Beans? If so, doesn't it involve a bit more overhead than simple Servlets-JSP-Beans (MVC)? I haven't worked with EJB, so I really don't know, just a thought. For a game, I think performance is a crucial factor. Working with servlets alone, we can't achieve separation of code and design that Herbert wanted, so we need to include JSP too. For the theme aspect, I think all that is required is a couple of CSS stylesheets (if only style changes). If the HTML itself needs to change, I think we must use XML with XSLT. Regarding the 3D aspect, i've written a minimal 3d engine in Java (so it would run in browsers without J3D). However it is pretty minimal in the sense that it supports only flat shading (no textures...) and almost no lighting. It's nowhere as powerful as DirectX or OpenGL, but maybe it could be a starting point. And the Timer, does somebody know how Yahoo displays the progress bar while the attachment is being uploaded? If that's known, we can follow the same principle. Regards, Senthil ------------------------------------------------- Sify Mail - now with Anti-virus protection powered by Trend Micro, USA. Know more at http://mail.sify.com Sify Power mail- a Premium Service from Sify Mail! know more at http://mail.sify.com |
|
From: Edward N. <cry...@gm...> - 2003-05-12 01:53:14
|
Hello James, Monday, May 12, 2003, 1:47:08 AM, you wrote: JH> Hi, I think that if we're going for a servlet approach we should go the JH> whole way and go for a three-tier system - this would have lots of JH> advantages: JH> o Automatically modular code, which encourages object-orientation and JH> helps people work together without treading on each other's toes. JH> o The possibility of writing a standalone client (this could incorporate JH> the 3d client. JH> o Lack of dependency on a specific platform such as a database or JH> webserver. JH> The downsides are that it's a bit more effort than doing it casually and JH> not many people are experienced in it. well i did a distibted systems programming module last semester, that wasnt as complex as this, but it was three tier. -- Best regards, Edward mailto:cry...@gm... |
|
From: James H. <ja...@is...> - 2003-05-12 00:47:13
|
Hi, I think that if we're going for a servlet approach we should go the whole way and go for a three-tier system - this would have lots of advantages: o Automatically modular code, which encourages object-orientation and helps people work together without treading on each other's toes. o The possibility of writing a standalone client (this could incorporate the 3d client. o Lack of dependency on a specific platform such as a database or webserver. The downsides are that it's a bit more effort than doing it casually and not many people are experienced in it. -- When a true genius appears in the world, you may know him by this sign, that the dunces are all in confederacy against him. - Jonathan Swift |
|
From: Edward N. <cry...@gm...> - 2003-05-11 19:07:51
|
Hello spacecombat-develop, On the subject of Java 3D, it can run in applets, its no extra effort. You can include Swing with it AFAIK but I have only used AWT just as it was quicker to write. Blackdown have produced a port for the Linux kernel but this port is only available for the i386 architecture, again not a complete Linux port for the many processor types which it supports. The Java 3D API is available for the following platforms, not necessarily produced by Sun: a) Java 3D for Windows (DirectX Version) b) Java 3D for Windows (OpenGL Version) c) Java 3D for Solaris/SPARC Each of the above API's are available from http://java.sun.com/products/java-media/3D/. The blackdown version for linux does run very slowly. I did not get a fps rate, but going by the implementation of java which they did it is aproximatly ten times slower than the windows runtime. It is not hard to install though. But i beleive 95% of the surfers do not have, or have not installed Java, so it would be asking a little too much of the general user to try and use Java 3D. Since no code has been written, how about making 3D an optional interface? -- Best regards, Edward mailto:cry...@gm... |
|
From: Kahless <ka...@sp...> - 2003-05-11 16:48:31
|
Hi, I've added you all to this developer list. Just to make things clear to everyone .. There were not written any code yet for the new (java servlet) version of SpaceCombat.. So we first of all have to think about the basic organization. On my request at sourceforge replied 5 people .. so we have 6 developers: Scott Goble James Harlow Senthil kumar Jide Orija Edward Neville and .. of course myself .. Herbert Poul I have not much experience with the development of servlets, but i guess this shouldn't be a big problem. On the technical side i want from SpaceCombat the following things: - speed (using java servlets instead of perl/cgi should solve that) - HTML interface where user can choose language and theme (so we have to seperate code, design and text output) - nice object oriented design of the code ;) Now to the content of the game... I think most parts of the old game design wasn't that bad but there are a lot of things which should be changed from the beginning. e.g. a variable amount of resources, instead of only metal, energy and population .. and that players can design their own ships... (putting some sort of equipment into slots of a basic shipdesign) Another thing would be using a 3d starmap instead of the old 2d :) The question is, if it would makes things too complicated for the player.. or the interface too complex. How I think we should start working .. First of all we should discuss how we'll organize the project (and how we seperate code, design and text ..)... Afterwards we should define how the basic system works (user interface, the resource system, ship system, etc.) Another important thing is the ticker.. I would like to make the game realtime .. but i have not really an idea how that should work. Maybe when some action is done (like building a ship) we can set some sort of Timer which is executed when the time is over and actually builds the ship. Does anyone have better ideas ? :) thanks, Herbert Poul P.S.: If you want we can also discuss things on IRC. I'm most of the time in the euIRC network (e.g. irc.euirc.net ) channel #spacecombat |
