PDA

View Full Version : Asphyre is now on Google Code!



LP
18-08-2012, 10:18 PM
In addition to being updated to version 2.0.4, containing few minor fixes and additional Image-To-ASVF example, Asphyre is now also hosted on Google Code (http://code.google.com/p/asphyre/).

What this means is that you can always have the latest working version of Asphyre by updating from its SVN Trunk (http://code.google.com/p/asphyre/source/checkout) (a feature that has previously been requested on several occasions).

WILL
18-08-2012, 10:45 PM
Great news! :)

LP
19-08-2012, 02:12 PM
In the packages there were some DirectX headers missing, so I've added the sources to the trunk and updated download packages. Sorry about that. :(

WILL
19-08-2012, 05:21 PM
So will others collaborate more directly with the development of Asphyre Sphinx now? Or is this strictly for access purposes?

LP
19-08-2012, 06:27 PM
Well, if someone wants to help with Asphyre development, I can always give commitment access for the trunk. ;)

The current development plans include better support on Mac OS and mobile platforms. There were plans to add OpenGL-based provider for Linux, but I've had some rough time running FPC/Lazarus there, specifically on Unity-based desktop, where Lazarus crashed when trying to change editor colors or the combo boxes become unresponsive and you could not exit Options screen without restarting application. I'm going to give it another try sometime later this year.

WILL
20-08-2012, 04:48 AM
It's always those tool glitches that really hurt development huh? :/

I am curious though, what kind of numbers do you get for downloads of the latest versions of Asphyre? Are you able to track just how popular a specific version is or how many are downloading it?

User137
20-08-2012, 08:59 AM
Are you able to track just how popular a specific version is or how many are downloading it?
You can track downloads in google code, i will let you search it ;)

Same doesn't seem to apply for SVN downloads. However cool thing is ability to track exact code line changes per each revision.

LP
20-08-2012, 11:24 AM
I am curious though, what kind of numbers do you get for downloads of the latest versions of Asphyre? Are you able to track just how popular a specific version is or how many are downloading it?

According to statistics, in previous month, there were around 1276 downloads of Asphyre Sphinx 2. There were some downloads for earlier versions as well, but considerably less. All these downloads, along with forums, used roughly 20 Gb of hosting bandwidth. This usage pattern has been similar since last November. For instance, in March, Asphyre Sphinx 2 got 1289 downloads.

I guess this number of downloads is not much, considering that above numbers might not be unique users and some people might even download the library several times or even different archive types (e.g. zip, rar). However, there was a demise of Delphi/Pascal users since few years back and only since last year there was some increase due to release of Delphi XE 2 with 64-bit and Mac OS support.

Eric
21-08-2012, 08:38 AM
The trend for Pascal seems to have reversed in 2009-2010 (with D2009 being the first version "usable" since D7?), but this wasn't really noticeable until end of 2011.
(ohloh does real-world activity stats on open-source projects, it's a better trend indicator IME than TIOBE f.i.)

https://www.ohloh.net/languages/compare?l0=pascal&measure=commits&percent=true

That said, if the commits went up, the stats for new lines of code added didn't went up in the same fashion, indicating that the activity is probably centered around bug fixing and porting the code to the new compilers/platforms. So not exactly a revival, but better than letting the code rot away!

PS: I registered your Google Code SVN in ohloh, its commits & code will now contribute to the Pascal popularity statistics ;-)

LP
21-08-2012, 01:22 PM
Although commits could indirectly imply the popularity of Pascal, I don't think it's an accurate measure as they are many more closed-source projects that are not being monitored.


The trend for Pascal seems to have reversed in 2009-2010 (with D2009 being the first version "usable" since D7?), but this wasn't really noticeable until end of 2011.
This kind of proves my above point. Delphi 2006-2007 was a huge jump since Delphi 7 both in language and IDE features. Delphi 2009, 2010 and XE weren't really big improvements, besides compatibility annoyance of forceful introduction of Unicode. Open-source trend in 2009 could be mostly due to delayed reaction to free versions of Turbo Delphi and growing community of FPC/Lazarus.

Eric
21-08-2012, 02:16 PM
Although commits could indirectly imply the popularity of Pascal, I don't think it's an accurate measure as they are many more closed-source projects that are not being monitored.
Yes, but activity on those closed-source projects is unlikely to exhibit trends completely opposite to those of open source on the same language, while TIOBE f.i. can be fooled by synonyms (a rise in "Delphi" fro TIOBE could actually be attributed to problems at the automobile parts manufacturer).


Delphi 2006-2007 was a huge jump since Delphi 7 both in language and IDE features.
Yeah, but they were dire in terms of usability, not as dire as D2005, but dire still, and those new features just weren't stable enough until D2009 (and even then, generics in D2009 are problematic, both at the compiler and RTL levels). The stability of the IDE was also not quite there.


Open-source trend in 2009 could be mostly due to delayed reaction to free versions of Turbo Delphi and growing community of FPC/Lazarus.
Lazarus is certainly showing activity, and the top individual committers revolve around Lazarus or FPC, though Lazarus itself is by far the largest FPC/Lazarus project, and it shows different peaks, f.i. https://www.ohloh.net/p/lazarus

As for real trendy languages, the winners are obviously Python & JavaScript, all the others seem to be either stable (like Java or PHP) or losing ground (C, C++, C#, Ruby...), just to place things in perspective.

phibermon
21-08-2012, 04:25 PM
I'm not sure how relevant it is to talk about the popularity of pascal in relation to other languages. If it's in terms of an open source project and attracting the largest number of users/devs as possible, then why choose Pascal anyway?

Given that the use of other languages such as Java effectively eclipses Pascal (pun intended) the only justifications I can currently think of are :

* Because you love the language

* Because of a large existing code-base that would be uneconomical to port to another language (including the use of specialized third party pascal code)

* Because you intend to write something big and believe the strong typing of Pascal will reduce your bugs in the long term.

--

That said I'm delighted that Pascal use is going up again!

LP
21-08-2012, 06:07 PM
I'm not sure how relevant it is to talk about the popularity of pascal in relation to other languages. If it's in terms of an open source project and attracting the largest number of users/devs as possible, then why choose Pascal anyway?
If you talk about Asphyre (remember, it's Asphyre thread), then the decision choosing Pascal (Delphi) was made almost 12 years ago. ;)

If you meant why keeping the project written in Pascal? Because it is too large to be ported to other languages, which would involve enormous effort and there might be better alternatives already. For instance, if you would use C#, then XNA would be a place to start. Besides, with 64-bit platform provided by Delphi and FPC/Lazarus, Mac OS and iOS, there is no reason why would you have the need to switch to a different language/tool.

Asphyre itself, which was initially called PowerDraw, was started based on motivation from experience with DelphiX. First, it was a plugin for DelphiX to provide hardware acceleration for alpha-blending and rotation effects, but as DelphiX started to age with its outdated DirectX 7 support, Asphyre (PowerDraw) was rewritten to be a standalone package using DirectX 8 back in 2001 (http://web.archive.org/web/20020123022023/http://turbo.gamedev.net/powerdraw.asp). Indeed, the peak period of highest activity in Asphyre community are 2007 - 2009 years; you can see this by the number of posts around these times in our previous community forums (http://dev.ixchels.net/forum/index.php).

WILL
21-08-2012, 06:57 PM
Asphyre has come a long way hasn't it? :)

BTW what are the odds that we will eventually see a new version codename for Asphyre? Not that the current one isn't cool enough or doesn't go it justice, but it seems that we've had Sphinx for some time and I do personally like the idea of eventually getting a whole series of codenames for future versions down the road as it continues to grow and mature with the developer tools it will be used with.

phibermon
21-08-2012, 10:10 PM
If you talk about Asphyre (remember, it's Asphyre thread), then the decision choosing Pascal (Delphi) was made almost 12 years ago. ;)

If you meant why keeping the project written in Pascal? Because it is too large to be ported to other languages, which would involve enormous effort and there might be better alternatives already. For instance, if you would use C#, then XNA would be a place to start. Besides, with 64-bit platform provided by Delphi and FPC/Lazarus, Mac OS and iOS, there is no reason why would you have the need to switch to a different language/tool.

Asphyre itself, which was initially called PowerDraw, was started based on motivation from experience with DelphiX. First, it was a plugin for DelphiX to provide hardware acceleration for alpha-blending and rotation effects, but as DelphiX started to age with its outdated DirectX 7 support, Asphyre (PowerDraw) was rewritten to be a standalone package using DirectX 8 back in 2001 (http://web.archive.org/web/20020123022023/http://turbo.gamedev.net/powerdraw.asp). Indeed, the peak period of highest activity in Asphyre community are 2007 - 2009 years; you can see this by the number of posts around these times in our previous community forums (http://dev.ixchels.net/forum/index.php).

My apologies, I didn't make myself clear. I mean in relation to discussing the current popularity of Pascal in terms of active projects (as others were) not Asphyre directly. In fact I was trying to point out it was a mute point as a polite subtle hint to the fact that it was an Asphyre thread ;)

I'd dearly like to see greater adoption of pascal for commercial game development and I believe that it's mature projects such as Asphyre that offer people feasible options to do just that.

Indeed I would be an avid user of Asphyre myself (as I recognize the elegant design of the current itteration) had I not chosen to focus my efforts on exploring engine design.

Keep up the excellent work! hopefully the new home on Google Code will also prompt more people to contribute to the project

LP
22-08-2012, 01:34 PM
Asphyre has come a long way hasn't it? :)
Well, it seems that one needs 10 years (http://www.selfgrowth.com/articles/10-Year_Rule_to_Become_an_Expert.html) to become master at what he does and it can be seen that Asphyre has converged in past 4 years to use multi-provider approach (which was once attempted roughly in 2004, but was not pursued further), allowing it to support multiple platforms and latest technologies such as DirectX 10 and 11 with Windows WARP (http://msdn.microsoft.com/en-us/library/windows/desktop/gg615082%28v=vs.85%29.aspx) device, without being limited to OpenGL only. I think after these years, in its current state, it is very polished, especially in regards to its runtime library (vectors, matrices, quaternions, colors, helpers and so on).


BTW what are the odds that we will eventually see a new version codename for Asphyre? Not that the current one isn't cool enough or doesn't go it justice, but it seems that we've had Sphinx for some time and I do personally like the idea of eventually getting a whole series of codenames for future versions down the road as it continues to grow and mature with the developer tools it will be used with.

Indeed, Sphinx has been the codename for latest versions of Asphyre for quite some time, which defined the multi-provider approach, perfected after short-lived Asphyre Casual. As there are no plans to drastically change the approach Asphyre uses, it may continue to use Sphinx codename for next few years. Current plans are to support more platforms, preferably including Mac OS and Linux with FPC/Lazarus (in addition to Mac OS through FireMonkey in Delphi), better support for iOS and support for Android.

The above is also motivated by our experiments, where we could port our game Aztlan Dreams to iOS (partially as we yet have to create Asphyre Sound for sounds and music ;) ), and it ran very fluently on iPad 2, even when using multi-core AI!