.
Annunci online

All that comes in the mind of an italian guy moved to california
TECNOLOGIE
20 settembre 2009
Diary: Technologies: AROS: Fall colors over the MESA...
Want to begin this article with a sad note: at the end of August my actual company decided to downsize and to change its location, and am among the laid-offs, so right now most of my effort in front of this keyboard goes in finding a new job in order to pay the bills and keep the apartment; hope to be able to dedicate time to my advocacy back soon (that of course means i will be employed): please cross your fingers...

[update in november: still no job, will let you know of anything].

The month of October looks like has been a bit unfortunate for this blog: several times, trying to look at the content, i was greeten from a white page: by the way the black-out (or white-out i might say) affects almost all the ilcannocchiale blogs: it is the second time that I sent notification to the webmasters about this and lately am starting to consider to move this blog in a more affordable platform, maybe even self-hosted.

But let's go back and talk about AROS. Right now the most important news in AROS front are the following ones:

As said last august, Krysztof "Deadwood" Smiechowicz ported version 7.5 of MESA to AROS, for now in software rendering mode only. Latter in the month he also ported GLU,GLUT and added a revised version of SDL that can use MESA. The addition of MESA as library allowed  people, especially Matthias "Mazze" Ruster to port some new games on AROS; among the contributions we have now Block Out II, a 3d Tetris clone, Abuse, a well-known platform game, a first attempt to port Open Red Alert (that later has been removed due to a bad bug in the upstream that prevent enemy AI to operate), 3d Pool Billiard (running kinda slow due to the software renderer, kinda playable if you keep the window at stamp size).

September news is that Deadwood decided to accept the Gallium 3d Bounty. Krysztof already started to meddle with Gallium 3d at the time of the MESA port; he started getting decent results, as seen in this movie here on youtube. Following those embrionic attempts, Deadwood compiled a version of GLExcess that includes the nVidia Gallium driver and, compared to the old software-only compilation, the difference is really remarkable.

The latest additions to the already rich porting gallery of Deadwood are the OpenAL/alut/ogg/vorbis libraries. Those libraries, as said from deadwood itself in AROS-exec, are used in several open source games to provide the sound experience; their porting will open the door, together with the perfecitoning of MESA and Gallium, to porting some open source 3d games, such as Warzone 2100, AssaultCube and Cube2; and is known the plan to use the hardware acceleration together with Cairo in order to improve Wanderer responsivity and performance, but as far as i know is all still related to Deadwood progresses.

Talking about 3d and Mesa, last month the e-zine The Vague came out for AROS using the last MESA port of Deadwood; e-zines are well known to amiga users thanks to the demoscene; though i was hardly been a follower, remember an Italian one made called Infamia - at the time even Tadsince1995 used to cooperate with it as coder - and reminding the experience I started The Vague. The AROS version is in the same .lha archive with the Amiga PPC and Morphos version; once double clicked the icon a requester ask to start either in full-screen (for 640x480 resolution) or windowed (as i started it being in qemu at 800x600x24 resolution) ,whether to use sound effects and the music replay rate. Nikos posted a youtube video of The Vague running on AROS here.


The Vague volume 1 on AROS


Its been exciting to see some good demoscene art and nice music on AROS: i liked The Vague and am waiting for further numbers, and possibly even more e-zines and (why not?) some AROS demos too :)

Last August the code of the Poseidon stack finally got released under APL and built together with the nightly builds, making so complete the Poseidon Bounty, the so far most contributed bounty in AROS history. I tried the stack on my old laptop, where a known incompatibility with the USB SIS controller so far occured, with November 12 nightly the USB devices are still not recognized, by the way. Once i find a way to download OWB from the nightly (writing the URL and then use wget to download it seems a good one), will post the dump of PCItool from my old laptop in order to help the debug.

Icaros 1.1.6 is already on the works but, since this blog post is kinda late, let me talk about the version 1.1.5, that I tried both as iso image on qemu and as live CD on the old laptop. So grub worked fine, but when I tried to select my sis900 network card with DHCP form the network control panel and press the "use" button in order to test OWB performance, the system froze hard. Everytime.
In defense of the network control panel, by the way, doing the same with the latest nightly worked pretty fine, apart the missing OWB in the nightly - beside an OWB folder and a OWB icon in extras/networking - for don't know what reason; i understand that OWB cannot be added to the nightly both for weight and license issues but my personal suggestionis to replace the actual useless directory with a script that might download the latest version on RAM, for new installs.

And, talking about OWB, our Stanislaw Szymczyk , after a while where he took a pause from AROS development, released another new version of OWB. The Version 0.9.9 is downloadable, as usual, from the sand-labs web site and this time introduces, beside the usual flattening of several bugs, a loader, called StartOWB, that uses the artwork made by michaels and helps to track the long first loading (caused by the fontConfig cache building); beside that, now the main application menus are finally available with the right mouse button as for every other AROS application, giving it a standard interface. As far as i know,  the AREXX port in OWB is still missing, but hope Stanislaw will fix this soon.


The new OWB splash screen,designed by michaels

And then let's talk once more about the hidden hero of AROS, Matthias "mazze" Rustler: most of the software used on AROS has been ported by mazze, starting form several games (, then the unfinished port of Ignition Spreadsheet (that i still hope one day will be finished), LUA and its extensions, zulu and siAmiga - that are providing the base ground for several utilities, panels and give easy way for beginner programmers (like me) to produce something in AROS: the last fatigue of Mazze is the porting of Scout, the famous low-level tool for manage tasks under Amiga OS; undoubtely this might help to get rid of some annoying freeze of some shell windows or commodities, though for the well known lack of protected memory on Amiga OSes is still better to reboot the system as fast as possible.

Going further, any AROS user had to deal with the well known problems with our version of Dopus: despite being invaluable for file management, it has strong issues with advanced operations, such as extracting files, viewing pictures, even personalising buttons; this is due to the fact that this dopus port is pretty old (from 2000, if i remember clearly): now a new bounty  has been introduced in Power2people.org, for the port of the latest version of Dopus actually for Amiga OS-4 only; the bounty has been accepted by Neil Cafferkey and hope will improve this invaluable tool.

Last September once again Pianeta Amiga was held in Empoli, despite the fact that this year was held only at Saturday and the flow of people is really at its own minimum. This Year AROS was highly represented from, as usual Paolo Besser and, straight from the United Kingdom, Steven jones that brought its iMica platform. But here is the outcome according to Paolo Besser:

Thanks to Enrico Vitali and other well-known people of the italian Amiga community, the event has been amusing and interesting, but... well, I can't hide the mixture of bitter sweet feelings that left me only half happy. The good news, at least for Icaros Desktop, is that every visitor of the fair had the chance to see it in action, and most of them stopped at my table, asking for more informations or to see some demostration. I've also got many congratulations and many thanks for the effort, asking me for keeping up "the good job", but the bad news about all this, is that Pianeta Amiga this year lacked of visitors. Numbers had been fairly low (I've personally counted about 100 visitors) and the exhibitors themeselves were a tiny group that could comfortably stay in less space than the half area of the Palaesposizioni reserved to the event. "In the good old days you could see a crowd of people making long tails in front of the ticket booth - said Enrico while we were taking something to eat at noon - but today...", unluckily today's Pianeta Amiga loosely reflects the situation of the whole Amiga market. A tiny, fragile community which is hardly trying to keep its platform alive, even if the interest decreases a little more every year, and even if the whole IT market has completely forgotten the Amiga. So we have to thanks ACube, VirtualWorks, AmigaKit.com and everyone else that still persists, and spends money every year to organize a event which has just become a symbol, a tradition for the community, but which is practically perceived like a hobby, a nice meeting day with other people who share the same passion, even by people who professionally operate in the Amiga business.

It is sad for me to hear that despite the good news in the Amiga market (the SAM availability last year, the MorphOS on Mac Mini, icaros, iMica, Ares one, the settlement of the Amiga-Hyperion cause,etc., one of the last Amiga events is gone so unnoticed; despite my hopes, i think it is realistic that next year Pianeta Amiga might not be hold, if this is the trend; of course i expect thngs to change, but who knows....

Frustrations and disillusionment might hit even die-hard supporters. Many of the Aros-exec usual lurkers know Nikos: beside being a strong fan of the platform he is also one of the testers and one of the main bounty contributors ever, but in this thread he say that decided to take a leave from AROS, disappointed from the apparent lack of progresses, especially in the hardware acceleration section. By the way this happened just before Deadwood released its new MESA port and announced the start of the Gallium3d Bounty. The community reaction has been something you will rarely see in other open osurce community: almost everybody sent its own comment to nikos (including myself) expressing support; that convinced him to stay and luckilly for him, the new results on AROS seems to keep him busy...

Another AROS-friendly machine recently surfaced: provided by Pascal "Phoenixkonsole" Papara, the ARES one is an AMD Athlon x2 powered machine provided in a small tower case, bigger in size than Steven's Atom powered iMica; the graphic card is a performing nvidia GeForce 7200 with 256 megabytes DDR2  video RAM; phoenixconsole expects support from Gallium3d in order to enhance 3d capabilities; the machine hosts 1 gigabyte 800mhz DDR2 RAM (can be expanded up to 8 gigs) and can host up to four 5,25" devices; has a DVD/RW drive and a 3,5" floppy drive (foreseeing a catweasel support); it has a SoundBlaster Live as sound card and the network card is a Realtek RTL8139d that has a RTL8139 family chip and, quite interestingly, just today Hitchhikr released an rtl8169.device driver that in theory should support it. Paolone had occasion to test it with its Aspire One netbook and said it works nicely, adding another device to the netbook puzzle (right now still wireless and sound are unsupported on Aspire One).

Phoenixkonsole also plans to put in bundle with Ares One Icaros Desktop and a suite of programs for AROS including: the registered version of FryingPan and the incoming AROS Port of Cinnamon Writer: the latter one fills another software gap under AROS, presenting the first Word Processor available for the platform.


Cinnamon Writer is about to spice AROS


Now, if only Ignition port were finished we surely might had a decent AROS base office suite, together with MuiBase; the actual problem in finishing Ignition is non trivial. According to mazze itself in this thread on AROS-exec, it is depending from the actual order of "struct node":

Problem is that AROS has the elements of "struct Node" wrong on X86. This is supposed to be fixed with ABI V1. Unfortunately, Ignition makes *heavy* use of linked lists.
I could:
-continue porting when V1 ABI is released

-fixing list handling in Ignition which would be a lot of work which wouldn't be any longer necessary when V1 is out

-compiling me a version of AROS with the right order of Node. I could continue porting but I couldn't release the result until V1 is out.

All sucks somehow.


Further in the thread, phoenixconsole ask for how long might still take to Staf Verhaegen to finish the ABI 1.1; Staf's answer does not look too encouraging:

If I don't get help months to years from now. Problem seems to be that for being able to work on ABI V1 you need very good knowledge of AROS internals, Amiga OS internals and low level programming. It seems it is difficult to find people that have these capabilities and want to do it as their hobby activity. Additionally no glory or money can currently be gained by it.!!!

Staf is quite right: actually very few people have the required knowledge to help him in fix the ABI, and the actual AROS documentation, that should help provide that kind of knowledge is still kinda incomplete: lately even the quite famous "Linux lady" Carla Schroeder wrote two articles about the importance of documentation in Linux magazine: they can be found here and here, hoping to give guidelines for the actual coders, and maybe some non-coder to help therm in write it.

Simone "samo" Bevilacqua finally ported BOH to AROS, and it behave nicely even under qemu! BOH is one of the new generation commercial software available on the Amiga platforms (a little but, considered the platform diffusion, important avantguarde of a hopefully bigger production that include even Cinnamon Writer itself) and has been prepared for all the Amiga platforms so far: os 4, morphos and now even AROS. Despite some gitches with the sound due to the SDL libraries (and i expect things to imnprove thanks to the recent Deadwood work), the game runs fine even on qemu.


BOH for AROS



This coming year, God and Cash willing, is my intention to attend once more the Southern California Linux Expo in Los Angeles; will introduce the new features of AROS, the Icaros distribution, the new software and, if possible, even a native AROS box. Will write more about it soon; just mind that if somebody want to help me next february can get in touch with me at the email on the right side of the blog - the Get in Touch with Simone Bernacchia link, that I will repeat here, just remember to replace the sentences between square brackets.
TECNOLOGIE
19 agosto 2009
Diary - Technologies - AROS: the king is naked and being hit by a bus...

For the first time I start to write the article in english instead of italian in order to give higher priority to the latest happenings. Will try to keep it short, considered lately I have been kinda late in delivering my usual articles.

So here we are with new ports: Fishy_fis recently ported Dosbox to AROS: despite some issues with the keyboard it mostly works and now allows to use old DOS applications and games, and even with some level of inaccuracy Windows 3.1. Paolo Besser shown in its blog how he was able to run the old Word 2 for Windows under Dosbox. Of course this does not mean that since we have dosbox and J-UAE we should not do new applications: AROS runs in much more powerful hardware than the old Amigas and DOS PCs and new programs using its capabilities are not only welcome but desired: furthermore (and this is not the first time i say it) writing new apps for AROS, will make it available, after a short work of adaptation, also for other Amiga OS systems, therefore three time (approximatively) the users.

Since some time ago i tried to write my own network application in amilua, seems that other people discovered the flexibility of zulu, and some small utilities start to appear.

Yannick "Yannickescu" Erb built WHD Menu, an alternate WHDload launcher in amilua that interfaces with E-UAE, and it looks pretty good: it includes screenshots (either coming from icons if in a supported format or from a screenshot directory) the list of titles and a custom interface GUI to configure it. Despite some minor tweaks (some tweaking is required by the user both on AROS side and UAE side according to the WHDload setting and path on the UAE machine) and the well-known limitations of actual Zulu (as the inability to dynamically update lists) the application has a professional look and shows the potential of the technology. I really hope that Mazze will be able to include callback hooks and hopefully an integration with the Cairo Library soon, in order to have our own base development language for newbies and rapid applications.

Last week the Poseidon stack had been put on validation and all users invited to contribute with their own bug report. I tested the custom AROS build with Poseidon provided by Paolo Besser in both my laptops, the old one, an ASUS a1300 p3/900 with 384 megabytes of RAM and 20 gig and the new one, a DELL vostro 1000 with AMD sempron 1,5 gig ram and 160 gig hard disk.

The test in the old laptop was kinda disappointing:poseidon did not recognized my USB OHCI hub and therefore none of the devices I plugged into it (platon defined my stack very old and bugged) while in the Dell vostro most of the sticks were recognized (beside an old staples 64 megabytes one pre-partitioned in two sections). I know a bug report has been filled for the SIS USB controller and also hear that Neil Cafferkey, haveing the same controller, was trying to see if there was a way to fix that.

You plug your device and a requester window pops up and, if you haven't made it before,it ask to name the device for DOS use and other parameters: this might souind normal to Amiga OS USB adopters that were used to this way of handle USB devices but for me, that i mostly used USB devices on windows, was a different approach: is not exactly a plug-and-play, can be more defined as a plug-configure-once-and-play, somewhat more expected in some occasions.

Some remarks on documentation: first of all, the Poseidon documentation is not available directly with the stack; it has to be retrieved downloading the Amiga stack from Platon's web site and is in AmigaGuide format, therefore the use of AutoDoc Reader is recommended; second of all, the way it was written assumes that the reader have a medium/advanced knowledge of the Amiga OS internals: in the AmigaOS world and time it was developed this made sense, since the only kind of Amiga user around was the die-hard one that used and abused its machine and can also do some little hardware repairs on it; obviously the AROS version of the documentation might need a rewrite, since AROS brings new users, sometimes completely unaware of the Amiga OS way of doing things and some other time not completely aware, as me and a friend of mine.

Will make a real world example: this friend of mine got an USB ethernet card with the dm9601 chipset,  that Platon declares supported on Poseidon documentation. Beside the fact that the card was not recognized properly because that version was not supported at the time (fixed), but we had no clue on how to use the card to get online.

Turns out that the device that has to be pointed to the TCP/IP interface is loaded in memory (and this does not mean RAM disk, just RAM) therefore there is no .device file to point; the device name should be simply declared either in the interfaces file in ENVARC:AROSTCP/db/ or, using the network control panel in prefs, first create a fake device file (like an empty text file), name it dm9601eth.device and save it in DEVS:networks/ to make it point from the TCP/IP interface.

That is a pretty singular Amiga way to handle some devices and is kinda unknown outside Amiga world; once again, considered that AROS aims, willing or not, to be the easiest gateway for newcomers to the Amiga World, it is my personal opinion that the Poseidon guide file should be provided with AROS and also re-written to ensure newcomers might be able to use it and configure it.

Beside this, there are also improvements on the part that requires Poseidon to boot from an USB stick; last August 4th, in the developer Mailing List, Chris hodges stated that:

Poseidon is now available at boot time by using the "enableusb" kernel parameter. However, as fat.handler or the cd filesystem are not available inside the kernel, booting from a fat formatted stick or CD rom are still not supported. Booting from SFS/AFS formatted stick or drive should be working, but I didn't test that.

Therefore we are in a situation where AROS COULD boot form an USB stick, problem is the stick is not seen from HDToolbox. Fortunately installAROS can be instructed to see the device in the same way that I explained before for the network NIC, writing the device name in the "device" text area if the "wipe disk" option is selected -  NOTE: I did not try that, since have no spare USB sticks to use, so please do not try to use this option, because i dont know if it might wipe your hard disk instead; this unless - as usual - you know what are doing and have back-ups handy.


Paolo Besser released the 1.1.3 version of the now well know main AROS distro, Icaros Desktop. As most of you probably already guessed, the most important feature is the inclusion of the Poseidon stack, but i would also like to point little contributions like the OWB quick handbook made by Nikos, as usual a new batch of system build files (from July 31st) and the inclusion of the LiveUpdater in the distribution.

[troubles with qemu, best fit and USB - and an annoying editor bug]
And so i decided to switch my old QEMU virtual machine with the old one. At first i found out that launching the new .bat file from a different position made QEMU exit with every selection I made form GRUB so i moved the new virtual disk and replaced the name in the batch file, and then it worked. Still, if i use QEMU and select a "Best Fit" option in GRUB, QEMU will close; since my screen is a 16:9 and the 1024x768 option is not feasible as window, i resorted to edit the file boot/grub/grub.cfg to add 800x600 modes, and found an annoying bug: deleting something in the middle of a string causes the first character in front of the cursor to overwrite all that is from him and the end of the line; had toe resort to press enter in front of all parts that i needed to modify to reduce the damage; hope this will be fixed on the coming builds because the fact that the main AROS editor is broken is kinda annoying.


The new annoying editor bug

Another thing that I wanted to check was whether Icaros in QEMU can handle USB devices thanks to Poseidon, so i looked in the net for a tutorial and found this wiki from the Slackware project where explains how to mount USB devices on it, well not exactly user-friendly but not even too complicated: first it is needed to switch in the monitor mode using CTRL+ALT+2, write the following command line:

usb_add host [vendor_ID]:[product_ID]

and switch again with CTRL+ALT+1.

In order to retrieve the vendor_id and product_id codes, since windows xp has no lsusb comand, I found this interesting freeware called USBDevView, that will show all usb devices installed in your system and all related data.


The USBDevView utility and, circled in red, the Vendor_ID and Product_ID codes required to mount USB devices in QEMU

Once i obtained the codes i tried to add my device in the way that was explained in the wiki, and had no feedback from inside QEMU. Now i dont know whether is poseidon fault, QEMU fault or maybe should have done something more than mounting the USB device, but my QEMU starts already with the -usb option and so i expected to be able to mount devices at will. Am accepting suggestions.

And, together with the very close completion of the Poseidon Stack, another interesting utility is about to be released for AROS: coded by Michal "rzokol" Zukowsky, SCANdal is a graphical fontend for Betascan (that is an Amiga OS port of the XSANE drivers from linux); is already gone out for MorphOS and will be out soon also for Amiga OS 4; thje actual development in the AROS side is actually stopped due to another Zune bug in displaying more than one radio button at once, hope Michal will be helped and find a workaround for it.

"Steril707" some time ago started to experiment using Rob's cairo port and see if it was able to take out something interesting from it: the result is what he called "Shotofop": a simple but effective graphic application that allow to do basic operantions such resize, crop, rotate paint and select parts of the paint. The first verison is quite primitive and it uses the Adobe Photoshop toolbar (but of course new and more original button will be used next); it also supoprts a limited number of layers. Steril plans ot implement partly the PSD file format too. I suggested him to get in touch with the author of SCANdal and find a way to make the two applications talk,it might be interesting.

Krzysztof "Deadwood" Smiechowicz ported latest MESA version 7.5 to AROS in shape of mesa.library; since AROS still have no hardware 3D support (delegated to the Gallium 3D bounty, and is still low - please donate!), all the rendering is actually done via software. So far GLU and GLUT are compiling but missing functionalities and incluided as dynamic libs, and SDL implementation is not done yet. In the past Deadwood ported the Eternal Lands Client embedding an old MESA version, now with the release of MESA as a library that will no more needed.

Not so good news from the Kickstart Replacement Bounty phase 1: i went to know through IRC that Greg "bheron" Casamento last month broke its leg and now is obviously focusing more on recovering its health; me and the AROS community wish him to get well soon.

Last month Amiga OS 4.1 has been reviewed from Thom Holverda of OSNews.com. An important preamble is that Thom started its informatic experience on BE-OS and Mac and is an active Haiku-OS supporter, therefore is mostly unaware of the "Amiga Way" of dong things, like our winndow system works, how libraries works, etc.

So the result is something that you can compare to a newbie introduction to the world of Amiga, and give what they call a fresh look on the actual platform state.

I might also say that he had some spot-on observations on things that I would like to see in example on AROS,IF one day will be decided to use draggable screens in Amiga OS style (Kalamatee and Rob actually dislike them btw):

What AmigaOS needs is a few default screens set-up, and the ability to drag and drop windows [and icons or files too when the screen contains a single app on my opinion - nda] from one screen to the other. Currently you have to manually configure windows to appear on certain screens, and while this is useful for running, say, old Amiga games, it's not very user friendly for new users like me. The feature has a lot of potential though, so I hope the AmigaOS developers can capitalise on it more in the future.

I am well aware that the commercial Amiga-like OSes are more advanced than AROS where i meddle almost daily (see iconification, interactive pointer and show as list),but some basic things remained the same, such as the way Amiga OS handle its windows, not clicked by front, or the need to update the workbench manually to show drives and files, and those are among the things Thom did not liked; while personally on the window handling topic I feel at home better on Amiga OS style (obviously), i sincerely miss the automatic update and snapshot in the windows icons and positions (AROS support automatic update, but not for FFS so far), and am not alone:

the file manager also doesn't auto-update its contents; you need to manually update a folder if you downloaded something new into it. There are 3rd party utilities that fix this problem, but I'd prefer it something as basic as this is part of the default installation.

...

Speaking of windows, the AmigaOS seems to have a more persistent problem with retaining window sizes - almost every application refused to remember window sizes, which really starts to get on your nerves after a few days.


The final conclusion Thom draws are good but not exactly the most positive around, and is not the first time those are reported on an Amiga board or blog:

The AmigaOS is cool. It's fun. For most of you, it will be a whole new world of technology to explore and play around with. It's also a well-implemented world, with a logical file system, flexible file layout (you can move everything to everywhere, seemingly), cool features like the draggable screens, and lots of other stuff. It's also remarkably configurable, and given more time, I would've loved to explore more of the innards of the system, to really be able to use the system to its fullest potential.

However, said fun and coolness comes at a massive price, and this time, I'm not talking about the price of the soft and hardware. Despite the lipstick the developers put on the system (in the form of transparency and other fanciness) it's still very clear that the AmigaOS is a relic, a thing from the past. The application portfolio is outdated, lacking, and incapable, there's no protected memory, and many configuration panels are overwhelmingly difficult to understand and use.

AmigaOS 4.1 just didn't let me in. It's like being invited by a friend to a party where you don't know any of the people there. Your friend promises to remain by your side and ease you into the group, but once you arrive, your friend wanders off into the crowd, leaving you by the sidelines. And the group of people have known each other for 30 years. And they're catching up to 30 years of shared history. And they really aren't interested in newcomers - this is a reunion, not a party.

Remember: those are the conclusion of somebody that used to get to know the world of computing starting form a different side of the fence; this means that many things and approaches that we Amiga/MOS/AROS users take usually for granted are approached with a different mind orientation, might take the example of somebody that started driving motor vehicles using automatic gear and instead somebody that started using directly stick-shift, with of course Amigans among the stick-shift users.

Being a returning Amiga User, and having myself not been actively involved in Amiga progresses after 3.1 (my 1200 system is a 3.0 and had no chance nor the money to update the software), i myself had to re-learn many of the glitches introduced with os 3.5 and 3.9, that made a nice fast and efficient system such as the original Amiga OS 3 in a much less efficient and patched kinda blob. But despite that i still have the Amiga mindset and that helped me a lot in the past; i went to realize that without my original Amiga Background a modern OS-4, MOS or even AROS might have looked, if not as dark as a linux, probably much more primitive.

That inspired my main answer in the board:

Well,is hard for me to be objective when a fellow OS is involved; Thom as an outsider of the Amiga world expressed its concerns about AOS 4.1; so far, i got to realise that the Amiga OSes, including my endorsed AROS are made "by amigans for amigans" in paraphrase to the usual saying for linux.
What I mean is that, when i got interested in AROS in 2006 and tried the live CD, the first thing that made me fell in love with it was the feeling similar to the one of using an Amiga OS, in good and in bad: there were undoubtely flaws, but were *our* flaws, stuff we Amiga users had to live with everyday.

Like the workbench: as file manager has always been not the best option: Directory opus or filemaster has been Amiga user best friends since 1989 to help overcome those flaws,and still Amiga and AROS users deal with it either with the commercial old Magellan or the open source dopus 4 revisited, in example.

And now some of the Amiga desktop paradigmas and usability might look outdated to people coming from other systems, while people like me , used that all the time, actually feel comfortable with the windows that does not stack on click, allowing to focus on the main task and handling stuff in the back; but again is all matter of perception and habits.

I am glad that after many years of inertia things started to move in the amiga world again, but the problem is there is a lot to catch up and so far good old Amiga OSes are now a niche market for aficionados and a hobby; and looks like it might stay like this for long time, and not aspire to be more until many of the flaws are catched up, though i have good feelings about the netbook market...

At the end my personal opinion is: if you never used an Amiga OS or like and want to have a taste of it but have no money to spend, the first answer is try AROS, considered it is free and runs in most x86 hardware (and in virtual machines too); then, once you got used to it, if you like it, you can go the next level and buy a SAM for AOS or EFIKA for MOS too, according to tastes.

So, as you can see i endorse in part Thom opinion: of course I would like to see addressed the main problems of Amiga OS/MOS/AROS, epsecially for the usability part; i also pointed on the factr that right now a real Amiga OS or MOS are hard to reach for any hobbyist with little money to spend and proposed once more AROS as main low-cost  and low commitment gateway to the Amiga world, despite its actual incompleteness, that thank god is slowly being addressed (but still no interactive mouse pointer, damn!); furthermore the upgrade of Wanderer is the topic of this discussion in Aros-exec, where somebody also proposed (once again) to port Ambient from MorphOS,considered that is open source (GPL licensed) and much more powerful than the actual Wanderer; against this there are two main problems: one is legal but minor - Ambient is licensed under GPL - and the other one is merely technical but bigger - Ambient uses extensively MUI 4 classes and Zune so far supports only barely MUI 3.8. Steve Jones gave the hint that he might have a way to obtain the Directory Opus Magellan source code (once some licensing problems are solved), but even that is a non trivial port, considered parts of it are written in assembler.

And, of course, porting applications and filling the gaps in AROS is pretty matter of lack of developers: despite new blood lately came in, still very few people have the knowledge required to handle the core system coding, and despite the actual progress, still AROS is not yet fully recognized as one of the Amiga-like extended family members, like is shown in a thread in Amigapage.it, here, translated through Google.


An interesting happening of the latest days in the Linux world, about the disappearance (now solved) of the main mantainer of CentOS (and also SVN admin, domain admin and holder of all the monetary donations of the distro) gives insight to what i heard slashdot users call the Bus Syndrome or,in short, how many chances of survival have a project if one of the key mantainers or key developers is gone suddenly missing (or, as a figure speech, hit by a bus).

Now, in AROS there is little organisation as known, but still the figure of Aaron Digulla, founder of the project, one of the admin of the aros.org domain and of the CVS server - therefore the one who can administer CVS accounts - has a primary importance; is well known to the Developer's mailing list subscribers that might take an undetermined amount of time between the request for a CVS account to Aaron and receiving the account datas. But What if for some reason Aaron will come to miss? I wonder: is its position and the fact to be the CVS admin essential for the prosecution of AROS? Michal Schulz [that finished recently phase 1 of EFIKA port and now is working on the ARM port BTW] is another key developer and i already expressed here my concerns for when he will decide to give up AROS development, wishing for all actual core developers to try to documentate as much as possible; now also considered, because of some server upgrade problem, the CVS certificate expired and it has been impossible to build nightly for a couple of weeks, it is my opinion and advice that since Aaron in the last couple of years dropped active development in AROS and is actively busy in other projects, might be a good thing if he decides to give Admin privileges to some of the core developers as a backup move, to prevent any incoming trouble that might happen to ther project in case he might be unable to attend it admin duties anymore.

To finish, another interesting bounty has been rekindled last yesterday: since last year Bill Panagouleas' DiscreetFX acquired the sources for the Video Toaster suite, they also started a bounty for hardware-abstrtact and porting the programs including the ToasterCG suite on the modern Amiga-like OSes, including AROS. Since many of the programs,despite being made in C include parts written in assembly code (with exception of digipaint, being entirely written in assembly) that might be not a trivial work, but if done can surely help the actual lack of good video processing software on modern Amiga systems and also be used as base for new video applications.
TECNOLOGIE
10 febbraio 2009
Diary: Technologies: AROS - Happy new year with Fireworks show!!!


My A1200/030 still working fine despite three years of inactivity

Sorry if i was not around lately, I have been in Italy for the christmas holidays and also had my religious marriage there. Now that I am back to the states, is time to talk again about AROS.

While i was abroad,i was, when pssible giving a look at AROS progresses using my PC in italy: a celeron with dial-up connection at 56k, therefore i was unable, due to lack of bandwidth, to see the newer deoms posted on Youtube and Vimeo until I came back to the states.

I also been able, to my own delight, to put back on duty my old 1200 in order to retrieve some more material to post on the net, such as old .mod files and found out how, after three long years of total inactivity, is able to boot perfectly (unlike my Macintosh Performa 6400, that cannot see its internal drive), and had the feeling of a real Amiga system on my hands; this brught, of course, in doing some comparisons with my actual AROS system that am using right now.
Beside AROS is basically an AmigaOS 3.1 clone, now its usability is of course improved from its ancestor and, despite some bad system hand due to Dopus, or some keyboard shortcut not working as expected, the experience is pleasant. If somebody would be able to fix the glitches, it might be even improved.

I recently installed the version 1.0.2 of VmWAROS in my virtual machine: following the instruction i first installed version 1.0 and then, once I retrieved the update ISO, AROS found it and mounted it on the desktop (and this is already a good improvement) and I simply had to make the "VmWUpdate" script run (its icon is under my software/System Apps in the Amistart menu); it updated everything quite fast, in around five/ten minutes. I will make some more tests in the future, also considered that am preparing < a href="http://scale7x.socallinuxexpo.org/conference-info/schedules" target="_blank"> my speech for the SCALE expo next February 21st (1.30 to 2.30 pm PDT program track B).

2008 has been the best AROS year so far: important bounties like self-compile and SAM-440 port have been accomplished: the first one, in particular, as stated from me here[link] has also raised considerably the porting of other applications from other operating systems, together with the GTK-MUI library developer by Kalamatee for E-UAE; furthermore, the development of VmWAROS reached its first milestone and now, at 1.0.2 version has bvecome even more stable and mature getting its own characteristics such as NewIcons (in a special verison for AROS made by Ken Lester itself), the PDF file reader PoorPDF and AmiBridge, a basic integration script system between UAE and AROS.

But the most exciting things begun to happen towards the end of the year: among them the ongoing port of Ignition, a nice spreadsheet program ported by Matthias "Mazze" Ruster

, then the port of mplayer made by Krysztof Smiechowicz starting from the Morphos port mantained by Fabien Coeurjoly; it is actually an "early beta" and has a quite basic GUI but has finally brought in AROS a decent video and audio support; then there is the OWB port by Stanislaw Sszymczyk that deserves to be treated apart.

And then we have the work of Michal Schulz, still ongoing with its own two bounties - the AROS Efika port and the USB mass storage bounty - its lalatest contributions on AROS before to dedicate itself full-time on the ANUBIS project: not too long ago he stated on the IRC channel that he had been able to boot an AROS CD throug USB CD-ROM: later in its own blog, Michal explained more deeply what was still lacking for an adequate handling of the USB CD-ROM on AROS:

Every modern USB presents itself as a SCSI conforming bulk-only mass storage device to the system. Therefore, in order to do any IO operations, I have had to do them in the old good SCSI style. Therefore, the mass storage class do contain a DirectSCSI method and exposes it to rest of the world.

What a coincidence! I have added a few-lines long implementation of HD_SCSICMD to the .device layer of mass storage. I have started AROS in QEMU and forwarded the USB->PATA converter with a DVD attached to it and AROS CD inside (I didn't want to test anything special, but it was the first CD I saw on the desktop). AROS booted and stopped. Then, the timeout errors appeared. A lot of them. A bit disappointed I've left my desk and did something else. Suddenly, the CD started to spin and AROS booted from USB CD device! Great!

A short investigation showed, that the CD filesystem used in AROS performs the INQUIRY SCSI command with a fixed size, whereas the USB protocol provides a much shorter data in return, which leads to a timeout. Then, it tries to get the full-sized TOC and such. Anyway, AROS team will either need to fix CD filesystem a bit, or I will have to give up 10 second timeouts in mass storage and introduce four pipes there: two of them with very short timeout (100 miliseconds or so) and two with the 10 second timeout.





So now AROS can mount and boot even from the majority of USB sticks; considered that the transfer protocol supported so far is USB 1.1, the best results can be obtained with SFS, that takes around one minute to boot; for some reasons both SFS and FAT are not supporting cache durinf the data transfer and, therefore, are quite slower.

Still from Michal Shulz, we have good news even on the EFIKA port: the last January 26th Michal put on line a first beta version of AROS for EFIKA here, followed by a second version here due to some problems from Open Firmware in retrieving all packages: Other than that, Michal is aware that this version has still big glitches and has been released for debug purpose. According to him The Efika 512 source tree, once upgraded the sources with the latest contributions, reached the size of 150 megabytes, including a complete GCC tolchain and some more utilities.

O1i accepted the Phase 1 UAE Integration bounty: the results achieved so far are quite good: UAE opens its amiga windows under AROS, the amiga pointer is synced with the AROS one ad the integrated windows, that now runs at almost full UAE speed, can be re-arranged, mopved and resized from AROS controls. At htis moment there are still things not working properly in the integration, such as menus, closing gadget, border gadgets and non-workbench screens. The border gadgets, according to o1i are the hardert to get implemented; plus at this moment is not possible to launch Amiga OS programs from Wanderer; it is required to have an Amiga OS window open(such as workbench or a shell prompt) to launch the programs.

Mmartinkna posted a screenshot of AROS running under Parallels Desktop in Mac OS X; From what Paolo Besser wrote in its blog, Parallels Desktop is offering a better experience compared to VmWare, also because sound is working there. For the best experience possible, Paolo advise to set AHI preferences as AC97 (both for Music and for Unit 0) and to set ConfigIP to use the RTL8029 card.

Kalamatee has recently uploaded in SVN the beta version of its RTL8168 PCI and Gigabit NIC driver, asking for testing in the forum. Having still several glitches the driver is not considered workling; he is also trying to fix further the ata.device,especially for what concerns the legacy compatibility list.

Stanislaw Sszymczyk is the best purchase made by the AROS community in the recent times: its skills as programmer were already highlighted from the accomplishment of the self-compile bounty and now are further confirmed from its new task: the AROS OWB port; the first screenshot posted before christmas, desipite being just a google page with wrong colors and basic fonts, made me get strongly excited; the further status updates came so fast that it was looking unreal: OWB can already adequately show complex pages like Google maps and Google docs.
And, as a bonus, Stanislaw, going on with the development, enriched the toolshed available for the porting of new software: the new SDL port fromthe mainstream solved the messed-up colors problem seen in the first test screenshot and the new version of SSL, the 0.9.8j got ported too.

Since the Bounty requirement list asked for a Zune Interface for the browser, and also that the browser itself might be used as a zune component, the solution adopted by Stanislaw, once he realised that for some reason the AROS Genmodule generated code was not liked at all from the Webkit code, to write a static library that interfaces between the C++ webkit code and the C zune code; being the WebView interface properly implemented there has been no further problem: the screen redraw is incremental and therefore there is no need to redraw the page completely as the Webkit port of Rob used to be; the only reason that made Stanislaw not use the graphics.library instead of SDL was the missing alpha channel support in the first one.
Stanislaw thanks in its page Jorg Ströhmayer for the good organisationof its code in the OS4 port; that helped him in build the GUI onm Zune; he also revealed that "has been a pleasant experience" code again under MUI: its own tests with RegisterClass and its limitations made him build its own personalised tabbed browsing class complete with closing gadget.
Writing the class took Stanislaw around one week of work and brought him of fix some Zune bugs and to better understand its inner workings; the final resulting class bring OWB to be able to do tabbed browsing. Another requirement for the bounty was the use of Datatypes; Stanislaw's opinion is that the best Datatypes use should be for image decoding: actually OWB use internally an object hierarchy belonging to the ImageDecoder class, that implements different libraries according to the file format to decode; therefore Stanislaw wrote a new class that uses datattypes to decode pictures but this class use is actually limited from the following factors:

  • Picture datatype doesn't support progressive image decoding
  • AROS datatypes.library has no support for DTST_MEMORY source type
  • PDTM_READPIXELARRAY metod of picture datatype returns a data array with 0 values in alpha channel, requiring additional time-consuming processing

Because of that, the class had to inherit all those restrictions and therefore the image got decoded only once its data is all in memory, using also temporary files for that (and of course missing the progressice decoding); this limitation set back the Datatypes support only for those file formats that are not supported elsewhere.
For the future Stanislaw plans to add UTF-8 character set support to AROS, to use Zune widgets for handling forms and to allow more applications to use the Browser class at the same time; the latter one, according to its blog, due to lack of results has been momentarily set apart.

The latest surprise is coming from James "Jahc" Carroll, author of WookieChat: just some days ago he ported finally its SabreMSN on AROS, filling another important gap in the software landscape of the operating system. James needed the port of Open SSL to port the application; the port has been made by Stanislaw Sszymczyk; some components such as codesets.library are not supported or are faulty but the port of SabreMSN shows as AROS finally is achieving some room and respect in the Amiga developers heart, also thanks to the hard work of its developers and supporters.

sfoglia
  

Rubriche
Link
Cerca

Feed

Feed RSS di questo 

blog Reader
Feed ATOM di questo 

blog Atom
Resta aggiornato con i feed.

Curiosità
blog letto 1 volte

Older Posts

Who links to me?
Get in touch with Simone Bernacchia

 




IL CANNOCCHIALE