Annunci online

All that comes in the mind of an italian guy moved to california
11 novembre 2008
Tecnologies:AROS:Anubis OS: Anubis, son of AROS and the "leap" legend
[Edit: sorry to have taken so long in publishing this since the italian version: been busy in my job and in real life.]

I would have liked to start this thread with some mythological tone, such as Conan the barbarian and similars, talking about heirs to the throne, heritages, traditions and other stuff like that: i sincerely admit i was tempted to do so.

I will instead start talking about a day, three weeks ago, when Michal Schulz posted a link on the #aros channel on IRC: this link.

It is nonetheless that the sourceforge page of what has announced this week from Damocles as Anubis OS, the latest incarnation inspired by Amiga OS, based on a Linux kernel and with a Graphic subsystem that, according to the Anubis mailing list, should be based on xcb,xlib and a mix of Glib/Gtk and Zune. Further details have been explained here.

In short, is their intention to do with Amiga OS what has been done from Apple in OS X: somebody already uised the term "carbonize".

From what both Michal and m0ns00n (Hogne Titlestad) told me, the main reason that make them start the project has been an essential steadfastness from other AROS developers towards more innovative goals: Damocles already complained about the conservative approach of many of the core developers. as i wrote here.
So, Michal, together with m0ns00n and Damocles started to define the base functionalities and requirements of Arix. Michal will begin tu put full steam on it once the Efika port will be done; task that is bringing along the comppletion of the USB mass-storage bounty, another important requirement.

In a following IRC chat session, Michal told me how Amiga OS seems to have an history of lost opportunities, both for management and markerting mistakes, but also for ego trips and "balcanization" of the resources. He thinks at ARIX (sorry i dont like the name Anubis OS too much) as the Amiga OS that should have been, with resource tracking, protected memory and all those features that first Amiga OS then AROS were unable/unwilling to develop for the known reasons.

Michal is a pragmatic person and for him the usual debate about whether if you use a POSIX kernel in an Amiga OS that ceases to be an Amiga OS is or not is over: he decided that it is and, i hope, wil prove it quite good in the future.

Anyway, has to be known that Anubis OS cannot be considered an AROS fork: technically AFAIK a fork happens when a project is duplicated and modified but, being Anubis based in a different kernel, figures as an indipendent project.

The announcement was taken not too good from the amiga boards: for the aforementioned reasons above and for the cronical lack of development that is afflicting all incarnations of Amiga OSes, a heavy FUD cloud spreaded around, like has rarely been seen before. The maini nconcern among other amiga users is the runaway of the already few developers towards a system that might not even reach an usable state, and the progressive death of the already existing systems.

Luckilly, open source projects hardly die: they might become unmantained until somebody else starts again to take care of it or integrates the code in their own project; luckilly for AROS this fate looks far from happening: the project look active as ever so far, and hoping this state will last.

And, like to show that AROS, despite the imminent departure of "Doctor" Schulz and the announced "pregnancy" (kitty is a female cat :P) is still vital, let's talk about the so-far spiritual heeir of the Doctor: Stanislaw Sszymczyk that, according to its job for the self-compilation bounty, now almost [edit: no more almost: done!!!] finished has achieved considerable result, such as:
 - non dynamic port of the latest versions of python and perl;
 - self-compilation at first in RAM, then under SFS of AROS in AROS;
 - extension of the POSIX layer compatibility adding instructions such as vfork(), wait() and waitpd();
 - DOS.library modification in order to allow soft-links;
Beside that, sszy also standardised the development packages in the variuous port of AROS under GCC 4.2.2 and included their compilation in the AROS build system in order to have a system of compilers and cross-compilers built for all targets.

Furthermore, Stanislaw gave a look to the sources of Netsurf and OWB; beside some minor issues, the browser can be ported quite easily starting from the OS4 port if the GUI was not a ReAction one; so Stanislaw is looking for someone (looks like o1i might help, and i want to mention ShinkurO due to its experience on workarounds for the zune bugs) to help hm in doing a zune GUI for it.
The OWB port looks a bit more complicated due to the fact that the Gladelib port is required under AROS and so far sszy was unable to do the port.

Another important update comes from the API completion report: Krzysztof Smiechowicz completed its review job that kept him busy since last april; according to its data the completion level of AROS is around 80%. This report is also the ground for a concrete roadmap to follow so that a version 1.0 of AROS will be available in a more human timelapse.

And is also a good step that finally in the mailing list somebody is starting to talk about the strong need for documentation. Krisztof Smiechowicz has already started in adding documentation on the ABI v1 here and hopefully some more important documentations will continue to come soon.

At last look also that somebody will finally take care of the printer.device in a laid back timing; about me i was looking and sent cdocumentation to the guy interrested to help himin the development, included the old Ghostscript port on AROS now no longer working [edit: seems a new port is on the way]; lets hope to see something soon.
23 settembre 2008
Diary: Technologies: AROS: Documentation Documentation Documentation brings Developers Developers Developers!!!
Hoping Ballmer did not patented its citation :P

Well, cause of my job duties this is not at all a new topic, but think is good to bounce it here. The Amiga Os 4.1 on Sam good new is a week old and probably, when it will be possible, will treat it here too but for now let's do business as usual.

One of the latest topic comments in my last article in the italian forum (translated here btw) was from Tadsince1995 and brought me to forward it and enrich it in Aros-exec:

(translated from italian:)
So, after Robert Norris, even Michael is going. The biggest problem is not only that they will not work anymore in AROS, but will be a huge disgrace the fact that so far those are the most prepared developers, therefore their own knowledge luggage, shared by very few little people, will be lost.
I (tadsince1995) always supported a "Documentation bounty" because there is a GRAVE lack of documentation for the low level part of AROS. Nothing about the USB stack, almost nothing on how to write drivers, etc. How an operating system can proceed if newcomming coder have no ideas in how and where to begin with?

Especially the fact that, obviously, people have other committments and there might be no time to do reverse engineering on the code, while having some good tutorial might save them lot of time.

And here is my enrichment:

I support fully the point of view of Tad: this is my personal request for Michael, for Robert, if he is still around, and for all low level developers: do not only leave the source available: do not assume people to have same skills as you: PLEASE DOCUMENT LOW LEVEL AROS, for the sake of future development following your contribution:

Amiga/AROS scene is too little to expect people to find enough documentation (outdated at most), and Amiga/AROS internals are too different from others to expect the inner workings to be understood easily.

Therefore the usual Open Source method to read sources in this case cannot apply fully.

[there was a withdrawn part to avoid discussions]

To the people still at work:

you know this project needs developers and, IMHO, if not a schedule -by philosophy-, at least it needs a post-it in the border of the screen where to step the fundamentals, in short: "a plan".

Forget one moment about ports: i think the system needs to be finished and polished. I have ShinkurO and others complaining sbout buggy MUI classes, myself in the verge to do a review of AROS for a magazine being a bit frustrated from the lack of visual feedback from the system as cursors or messages; LUA working but zulu not fuilly implemented in case somebody needs to write new applications, not only games (thanks mazze for your efforts in trying to build SDLbasic).

[end withdrawn part]

I feel emotionally involved with this project and , as Paolone, i think it has more potential than even eople here can expect. So, let's try to work in a synergistyic (collaborative) way and pave the street for newcomers as much as possible.

That is, my personal message at the community; i really wish AROS not to sink, i already said in past and will continue to say it in the future.

My personal contributions are actually in helping ShinkurO together with Tadsince1995 in translating its Programming guide for Amiga Oses and in writing a review of AROS for a magazine, even if both activities are going on slowly in the free time due to the aforementioned job duties.
11 agosto 2008
Diary: Technologies: AROS: There is still something i did not bothered to talk of?

(picture from ok-cancel.com)

I am not in my best condition recently.

I am a bit late in my work schedule, have an acid reflux that make me feel like having a frying pan inside my stomach, and that makes it difficult to focus on my stuff.

I just read the continuation of the goodbye therad of Dammy, ex moderator of Team Aros, in this link on Moobunny and i feel a bit discouraged.

Dammy decided to take out the frogs in the throat in a quite harsh way: will cite its words specifically:

1. Worshiping PPC as the second coming is retarded. It's a failed desktop CPU arch, get over it already. Yes, it's viable as embedded CPU SOC, but that isn't an Amiga by any stretch of the imagination. Dare I say, SOC is the anti-Amiga if I ever heard of one. Unless PPC can lead us to a $150 laptop, it's a dead end. Amiga community made a huge error and it's paying the price for failing to see where PPC was heading off to. If they can make sub $150 laptops, PPC folks should be easily be able to make $50-$75 desktop/server box. Let me further define this, I'm not completely against AROS being ported to PPC. I had finally agreed for TeamAROS to accept a bounty for an EFIKA port (and then SAM440 port) because it would open the door for more developers and possible funding by OEM It is just one of many steps that AROS needs to grow into a viable modern OS to support many archs as possible. ARM should be next to be ported.

2. Lets talk another retarded decision about worshiping 3.1 API. What really drove me away was the brain dead decisions in sticking with a elderly API. The only arch that AROS can reasonably stick with 3.1 is AROS68K because apart of that API has to do with Amiga gfx. Amiga was the fore front of OS development in it's time. C= would have changed the API by now had they survived. Deep down even you know that, I know that, everyone knows that except for the 3.1 blockheads. Yeah, AROS would have been harder for some to port 15 year old software but it's 15 year old software, get over it already. AROS needs to be here and now and ready for the future not living in the past before it was even born! Cargo cultism at it's finest. If I absolutely had to run a old Amiga program, I would run it in UAE. Bill Gates has won, AROS is crippled.

By the way, Dammy is scrupolous in point the NON-developers as guilty on the fully featured 3.1. API compatibility: probably lots of them are notalgic people that strive for a "full immersive" AROS experience at the level of the old Amiga os to revive the "magic" and wake up from a 15 year long nightmare.

In my own, i really hope that AROS will accomplish the full 3.1 API compatibility soon and also hope that the goal will be a new start, not simply the end of line: honestly i don't care if we will reach the point where API needs to be broken in order to add new functionalities as protected memory and old applications might need to be sandboxed [run in a sealed memory area in order not to bring the system down with them in case of crash: the name comes from the sand box in the kids playplace] to run.

Well, not too much things to comment about: the AROS world, so as the actual Amiga/MorphOS world is a bit sleepy, like a country town: a lot of stuff are homebrew and made under the hood: to see how the engine is actually built and things are progressing helps a lot to be subscribed in some developers mailing list.

Sometimes we have nice bursts: the debuto of the Sam440 version, the new nightly builds that now are supporting GRUB 2, guaranteeing a better compatibility and let the system boot even from SFS partitions not only FFS; the EFIKA port where Yakumo9275 and Michael Schultz are working on parallel; the interesst for AROS from the Natami project developers, which should bring finally at a 68k platform port for AROS, and i already see the bounty assigned to bheron, but have no idea who this guy is: info anyone?.

So, in this small town the atmosphere brings to clash egos and well estabilished habits, in  the form we already know such as allthe usual arguments, missed implementations and integral convinctions on fighting between the different Amiga OS-like incarnations.

I don't think to be immune from it, would be superbous sayng so, buit at least i try to have a better real world oriented mindset toward the topic.

Might be that both me and Paolo Besser (just become a father so my congratulations) are ,in different ways working and struggling to make AROS grow beside the simple hobbyist operating system, as some of its own developers think still about: we are convinced that the Amiga architecture still has a very noticeable potential, especially in some hobbyist sectors, as support tool: since not all uses of an OS require protected memory and other advanced features, there is a way to increase the userbase beside the ex-amiga user and occasional developer/lurker: think in example at HAM-radio, at music sequencing software with MIDI enabled trackers and other open source software as audacity.

As I said on this thread (in italian) on Amigapage.it:

I see interesting openings especially for those hobbyist fields in which people prefer to focus more on the computer as help than as an additional instrument to be handled and tuned apart from the rest: i remember some applications for HAM radio on amiga, some lighting software (and i remember sources were released some time ago) and for other DIY sectors where is not required to have a fully modern operating system, needs to be simple to be used and configured and, considered that viable alternatives for those people are, for now windows or linux: both quite bloated and/or difficult to be configured/maintained and both requiring quite powerful hardware and a good amount of RAM: unlike those, AROS looks perfect: is lightweight, can be used in fairly old machines and is easy to handle.

Well, beside this, what else I did not bothered to write on? About Rapid Application Development tools to help the development of applications, something which all amiga flavours but specially AROS are lacking: for AOS 4 we had the no longer supported ADV, so some developers are trying fo till the gap. One of the developers of AROS, gianfranco "Shinkuro" Gignina, asked about an opinion on this kind of tools show a partial disagreement:

I personally find automatic GUI creator tools as something discrete for beginners and professionals on deadline. In some other cases I consider the automatic GUI generation tools a sort of gambling, which I would not use for serius programming.... i might add some real life example as Visual C#, that, when you use the tools does links and instantiate all GUI elements in one method, something that, once referred to OOP programming, based on extension and inheritance looks at least obscene....

..and not to forget that on Amiga Programming, the tool code and the GUI code cannot be separated as in more modern operating systems as cited by Shinkuro in this other thread:

There is no complete OOP API set on Amiga, that has been done only the GUI so far.
The Frying Pan author, to make its own life easier and transfer code faster between Amiga OS flavours, created this:

I think the link above might be quite interesting but, since has been released under GPL 2 license, i personally have no idea whether there might be any problem in using it iside AROS, maybe can be used for applications, and not for OS components...

It is also not possible, IMHO, to divide the GUI development thread from the GUI usability thread, that are actually the pain-in-the-bottom of most open source projects; some recent and less recent studies and considerations here and here let us know how the base problem is sometimes coming from the open-source environment nature itself (this is mainly related to linux operating systems, by the way) and the fact that coders does lack the skills - a subset not easy to achieve - to build usable interfaces.
 It is not completely their fault: writing an usable application requires nontrivial efforts and sometimes an application has to be designed to be usable since scratch. Plus usability is not an unniversal rule, is something that is quite subjective: what is usable for a developer is non usable from a graphic artist and vice versa plus what might be usable for those two categories might be hard to understand if not completely usuitable for an average user...

In the closed source world software usability on an OS is ruled by interface guidelines (something that GNOME and lately KDE started to adopt too) and improved by tests and feedback from users: - for what is concerning the first one, related to AROS, Shinkuro proposed an Interface Guideline reference on the developer list to improve the existing 3.1 guidelines, with not too much success so far;
-the latter one instead is undoubtely expensive, therefore hard to be put on practice from an open source project if not backed up from some capital: lately i learned about the existence of an open source usability project:  openusability.org, maybe can be useful.

Another problem for the open source application usability is the subculture hacker environment where some of the developers grew up: a very closed environment, almost a sect in which the initiation ritual is to understand and master the operating system and its application, deep to the inner workings an ddo the leap from "luser" to "l337": that might give an hint about the still scarce desktop penetration of linux: those topics are part of the bread-and-butter of the linux haters blog, and also this older but still interesting article shows how the supposed "lack of respect for the fial user" might be reconducted sometimes in the well-known for amiga users ego clash and its consequences, not alone the "leaving the playfield bringing all the toys with him" bad attitude that killed or hamed badly old software projects, including poseidon and MUI, instead of give the community the possibility to mantain and make grow opening the source: i know tihs is personal decision  matter but sometimes the reasons to screw up projects are so trivial or inconclusive that might sound like tantrums.

I hope to give a nice help to developers willing to improve software usability in linking this usable software small handbook from joelonsoftware.com: not to be considered an all-in-one course but a good start to improve the development method.



Feed RSS di questo 

blog Reader
Feed ATOM di questo 

blog Atom
Resta aggiornato con i feed.

blog letto 1 volte

Older Posts

Who links to me?
Get in touch with Simone Bernacchia