There is an interesting and brilliant article on Morphzone about various solutions of approaching installation procedures into Amiga-like OSes and mainly into MorphOS operating system.“Is it really necessary to install a program?”. This is the question risen by the article, and it shows also new guidelines which could be adopted on future Amiga-like OSes. It is written by Doctor_MorbiusFP (alias prof.Fulvio Peruggi) and by programmer Tokai (alias Christian Rosentreter).
Professor Peruggi also donates the community two AREXX scripts which could help any user in literally “trapping” any installiation procedure into a directory-tree (different from the default one) that replicates System Hard Disk Partition on Amiga.
You can find full article here.
Rise and Fall of Installation Procedures
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
when it comes to installations i must say i have fallen in love with the system used in the gobolinux distro.
pop the new application into a version numberd folder under a folder named after the app. put a symbolic link next to the versioned folder, pointing to said folder and name the link current. presto its installed
wow, really? thats pretty cool
I’ll have to try that distro out.
the most nuts thing is that its started as a kind of rootless system. ie, being able to install and maintain a collection of applications inside a users home folder. while the rootless release is getting old and rusty it could even be used under cygwin! (and may be so again, some people are looking into cleaning up the rootless system).
still, some apps needs to be recompiled to work under this enviroment (mostly those that have hardcoded paths or similar linking, bad bad).
but one can find compile recipies on the site, grab that, a matching tarball of source and aim one of the scripts that come with gobolinux at the set and you should have a working app. you can even generate a binary tarball for future installs if you want to. and they allso keep a nice collection of compiled mainstream apps on their site allso.
hmm, it would be kinda interesting to have this kind of setup in a corp enviroment. compile and test then generate a “package” and spread it to the diffrent systems via install scripts.
It’s hard to beat OS X. You just drag and drop the icon. Uninstalling is just dragging the icon to the trash.
The icons you see are actually self-contained app bundles with all the files and resources inside them.
Or installer packages, or VISE installers (developers, I beg you: do NOT distribute applications with VISE installers! DO NOT DO IT! Use .pkgs! …please!
And don’t ask for root privileges unless you really really really really really need ’em.)
sure, its a nice trick. but its os x we are talking about then, not a linux distro
but how do you update them?
that method works if you have a small number of programs (well, in commercial systems, applications uses to be few, big and do about everything) but in linux even with a basic install you got +50 packages…
now, may be “synaptic” can be modified to show a two panned window, where you drag from the repository pannel to the installed pannel, and when you drop it, a message tells you are about to install a package and its dependences…
just an idea.
The article seems rather vague in places, doesn’t really understand usability and IMHO overlooks some important points.
Reasons that installers were developed:
– Compression. Installers existed for DOS programs, people seem to forget this so far. They existed because it was a good way to fit a program onto a floppy disk without requiring the user to have a copy of pkzip or whatever
– Copy protection. Only install if you get the serial code, etc
– Yeah yeah, the registry. Give me a break. Registry is a bazillion times easier for developers than huge INI files, which was basically the alternative. It also allowed for things like sensible (ish) file associations. If you look at how file associations work on Linux which does not use a registry, I dare you to tell me it’s simpler and easier to clean up. I’m not saying it’s worse, but I’m really not convinced the “file and re-gen cache” design is much better than the registry. It has its pros and cons.
– On Linux, s/installers/packages/ and you get dependency management. The less said about that the better
By the end of the article he’s proposing obscure error messages asking the user to manually unload apps due to a quirk of the Amiga shared library system – is this for real?
OK, so what’s the system for dealing with configuration files, static and dynamic data files, and separation between files which should have different privileges?
The standard *nix filesystem exists _for a reason_, believe it or not, it’s not just there to annoy you.
how many clicks were needed to get to the article????
several pages and new windows later.. the article it was a good read.
Maybe the link could actually point to the article in question.
I would like to clarify a few facts concerning my article. Otherwise people
may judge it from an incorrect point of view. It is important to realize that
(1) the article was not written for a large audience but only for users of
Amiga-like computers; thus
(2) only 1/4 of the article is dedicated to a very short historical overview
concerning all platforms; indeed
(3) the main goal is not to discuss the growth or elimination of installation
procedures on all platforms, but only
(4) the reduction of installation procedures to a minimum in Amiga-like
systems.
> The article seems rather vague in places, doesn’t really understand
> usability and IMHO overlooks some important points.
Please see (2) and (3).
> Reasons that installers were developed:
> – Compression. Installers existed for DOS programs, people seem to forget
> this so far. They existed because it was a good way to fit a program onto a
> floppy disk without requiring the user to have a copy of pkzip or whatever
In my first Amiga (1988) I had 5 Mbytes of RAM and no hard disk. I created,
with early packing utilities, a file that contained the whole OS. In such a
way I put the contents of two floppies in one crunched self-expanding file on
a single disk. At boot time the file was run and decrunched itself into a
RAM-resident reset-resistant bootable virtual disk. And then I worked like I
had a hard disk. This was before early PC’s reached the MS-DOS 640 Kbyte RAM
limit, and before pkzip and similar programs appeared. So don’t tell me that
compression is important, I know this very well. See (1) and (2), instead.
> – Copy protection. Only install if you get the serial code, etc
You forget that these protection schemes were changed and/or abandoned as soon
as developers and distributors were aware of the existence of many “next door”
pirates.
– Yeah yeah, the registry…
If you like it, take it and be happy. I don’t.
– On Linux, s/installers/packages/ and you get dependency management. The less
said about that the better
Yes, now skilled people can find good dependency management. But maybe this is
an index of the bloating grade of an OS… Don’t you think that it would be
better to have no dependency problems at all? In a small footprint proprietary
OS this is automatically guaranteed.
> By the end of the article he’s proposing obscure error messages asking the
> user to manually unload apps due to a quirk of the Amiga shared library
> system – is this for real?
Sorry, you have absolutely no perspective. The system discussed in the article
(PegasosII, a PPC based computer, and MorphOS, an advanced Amiga-like OS) is
now used by a small community of amateurs and hobbysts. Maybe it will evolve
and expand to a larger user base. It inherited some features of old Amigas,
among which the management of shared libraries (no quirks like you think).
Since we use many legacy applications, we need compatibility and must find
temporary solutions. If the system will survive, legacy will be less important
and the behaviour of the system will be changed.
if its gobolinux your asking about adamw, go check out one of the articles on the site: http://www.gobolinux.org .
it may explain it better then i can do.
Using a system like 0install makes instalation simply a non issue
http://0install.net/
Applications are never installed, only cached on the system.
I would love to see a distro based on this… something a little bit more frendly than b00t
http://web.archive.org/web/20041014185850/http://b00t.webstep.net/
Amazing how the issue of getting software off, and onto a computer gets so complicated.
In my first Amiga (1988) I had 5 Mbytes of RAM and no hard disk.
You weren’t the president of the World Bank in 1988, where you? Or the International Monetary Fund? No pay anyone sensible could afford that much RAM back in 1988.
512KB, now that’s a different story.
An Anononymous person wrote:
>how many clicks were needed to get to the article????
>
>
>several pages and new windows later.. the article it was a
>good read.
>
>Maybe the link could actually point to the article in >question.
😉
Direct link is right there. You can find it if you click “READ MORE” and read complete news about it here on OSNews.
😀
First link that you saw points into Morphzone site as a reference and netiquette regarding the site which host the article.
However here follows complete link. I posted it for those who arrive right now on OS News reading about this article.
http://www.morphzone.org/droffice/RiseAndFallOfIPs/Rise_and_Fall_of…
OK, so I understand that it’s written for Amiga/MorphOS users. But I don’t think the right way to improve MorphOS is to work around quirks (yes quirks!) in the way its shared libraries work. There aren’t _that_ many Amiga apps, so legacy issues just aren’t as big a deal as they would be for some other OS.
I don’t like the registry, but so far the alternatives Linux and MacOS X have come up with for doing many of the same things are either not convincingly better or are actually mini registries themselves. For instance copy a file and regenerate a cache – well, this works OK but there are quite a few subtle issues to work out. And you need fast filing systems that can cope with thousands of files in the same directory (but this was always true of UNIX). And the MacOS X file association system seems to rely more on luck and the user following kinda-but-not-quite documented conventions.
Sorry if I came off sounding harsh, but the first part of the article at least reads like an all too familiar “all other operating systems are so stupid, installers are stupid, nobody needs installers, just copy the directory!” type article.
> You weren’t the president of the World Bank in 1988, where you? Or the
> International Monetary Fund? No pay anyone sensible could afford that much
> RAM back in 1988.
>
> 512KB, now that’s a different story.
I’m sorry, you are in need of a history lesson.
I will tell you the story of 12 professors of the Department of Physics of the
University of Naples that, tired to share the processing time of an IBM
mainframe with the whole Faculty of Science, decided to buy 12 Amiga 2000
computers. We gained the following:
(1) A multitasting OS when PC’s did not know the meaning of that word, and
Apple MacIntosh’s only had inefficient task switching.
(2) 1 Mbyte of RAM already available on board.
(3) A complete TeX environment (AmigaTeX by Radical Eye Software) for creating
our scientific papers.
(4) Fortran 77 (ABFortran by ABSoft) for all our computing purposes.
(5) Plenty of games to play with in our free time…
Usually we worked launching at the same time the “Ed” (or “Memacs”) screen
editor for Fortran and LaTeX sources, the TeX compiler ready to be used in a
shell window, the TeX graphical previewer on its own screen, and the Fortran
compiler and linker ready to be used in another shell window. At first we had
the OS (AmigaOS 1.2) on a floppy and TeX and Fortran on another floppy. Hard
disks became usable one year later with AmigaOS 1.3, and I had low memory
problems as soon as I wanted to diagonalize a larger matrix. So I decided to
buy a memory expansion board as soon as Commodore made it available. It had
2 Mbytes already on board and empty sockets for other 2 Mbytes, and, yes, it
costed as much as the whole computer. But I needed it for professional
purposes, so I purchased it and soon started to be very happy (and created
the virtual disk environment described in previous post). One year later I
also populated all the empty sockets and upgraded to 4 Mbytes.
In conclusion I was there and was very active at that time, while most
probably you were not. So, please be sure of what you say next time…
There are tens of thousands if not hundreds of thousands of Amiga apps. Even if only hundreds of them are used today it is still a significant number.
> OK, so I understand that it’s written for Amiga/MorphOS users. But I don’t
> think the right way to improve MorphOS is to work around quirks (yes
> quirks!) in the way its shared libraries work. There aren’t _that_ many
> Amiga apps, so legacy issues just aren’t as big a deal as they would be for
> some other OS.
There are more than 20,000 Amiga applications and games. Even if 19,900 are
outdated (but we can run all of them within UAE), there is a non-neglegible
set of still useful applications.
Currently MorphOS is developed on a voluntary basis. It needs support and
sales to collect funds, so development is slow.
At the same time the Pegasos/MorphOS user community (I estimate two thousands
persons at most) wants to be useful in the evolution process, for instance
trying to solve autonomously some problems. Since at present time we must
still cope with Amiga’s legacy, my proposals were directed to other users and
programmers that would like to create temporary patches. I expect that the
MorphOS team will never consider them as targets.
> Sorry if I came off sounding harsh, but the first part of the article at
> least reads like an all too familiar “all other operating systems are so
> stupid, installers are stupid, nobody needs installers, just copy the
> directory!” type article.
Yes, Amiga people is well-known for snob attitudes. But it is not so, in this
case.
We have solved 50% of our survival problems since a more modern hardware
(PegasosII) was available. Another 30% of problems are solved by MorphOS,
but it must still volve, and much time is needed.
Anyway you should not think that, instead of amateurs and hobbysts, we are
a bunch of nostalgic fools. On my PegasosII I run Debian Sarge when I have
compatibility problems (e.g. I use Mozilla when the outdated Amiga browsers
are incapable to display certain web sites), and I also run MacOS X on
MacOnLinux when I need a higher compatibility with the external world (e.g.
MicroSoft Office). But, since I like lightning response times and simplicity
of use, I still work most of the time in MorphOS environment (e.g. I write
scientific papers with AmigaTeX, and solve symbolic mathematics problems, as
well as numerical calculations, with MapleV).
So, in conclusion, the hardware is not the original Amiga hardware, but we
are still blessed by the extraordinary flexibility that always characterized
every Amiga system.
>1) A multitasting OS when PC’s did not know the meaning of that word, and
>Apple MacIntosh’s only had inefficient task switching.
>(2) 1 Mbyte of RAM already available on board.
>
[Blablabla…]
No… Please stop that. There’s no way you can be taken seriously with such talks. I’m a MorphOS and am really sorry to hear such things
We know what the Amiga had that PC/Macs gained later but this has nothing to do with the original topic…
Btw, Amiga’s flexibility has its limits. We are (have) reaching it. One day we’ll have to drop compatibility (crap legacy). The sooner the better.
Leo.
I have to agree about dropping legacy support. The 680×0 emulator is great but let’s just stop there.
Focus on the here and now. Freepascal support is one small step forward. Maybe if some modern framework could be ported to help people develop some modern appz that would be great. Things like that would help I belive.
About the registry. This is acctually a good idea, but not in the case of the “Windows registry”. That’s FAR to messy. But the general idea is good.
> No… Please stop that. There’s no way you can be taken
> seriously with such talks. I’m a MorphOS and am really
> sorry to hear such things
Read everything carefully before telling things unrelated
from what I say. I have never said that a few performances
of the _PAST_ are important today. I only clarified those
facts because someone accused me to report false
information. But I’m old enough and lived those stories.
Today the only fact that matters is that Pegasos hardware
flexibility (not Amiga software flexibility, that you
erroneously cite, and that has no more relevance) allows us
to run MorphOS, all Linux flavours, and MacOS-X. This is
important for us because:
(1) if MorphOS dies, which is still possible, there are
already alternative solutions;
(2) if MorphOS lives, it will finally change, drop
compatibility, and become better.
So we are saying exactly the same thing…
Anyway, note that this discussion is absolutely boring for
most people here (I’m sorry for that). In fact, why should
they care for MorphOS and our negligible niche?
Their interest may be attracted only by Pegasos computers,
an open hardware that may free them from x86 dependency.
No offense on the Amiga front (hey, I had 3 of them myself, last one was a A1200). I just commeted on the (for the time) insane amount of memory. Yes, I knew the Amiga could cope with 8Mb. I remember paying shitloads for 1/2 Meg, and just thought that anyone paying for 4 MB expansion must be rolling in money back then.
I was a student back then, so $1000 was a fortune for me. I’m sure that the 4MB RAM set you back at least $2000 in 1988 dollars. That really is a lot of money.
Zenja – What you are doing is thinking that the Amiga was upgraded like modern PCs. You are correct that it would cost a fortune to upgrade the memory if you bought standard Zorro-II memory cards and plugged them into the bus. But depending on which model Amiga you had there were other options. In my Amiga 1000 I upgraded the chip set in pigbacked the memory chips directly to the motherboard memory – that upgraded me to 1 megs for about $150 CDN, then I designed and wired up by another 2 meg board and wired it into the auto-config space. That cost me another $150. So I was running 3 meg for about $300 CDN in 1987. Later I added 1.5 meg of static memory tied directly to the CPU’s socket in an experiment involving a clock doubled to the CPU but that static memory costed me big time.
People upgrade their machines in ways no-one would consider doing to a modern PC, even if it can be done at all.
I never had the problem of installing apps on my hard drive on my Amiga because I only had my first hard drive [a whole whopping 65Mb!] when my use of the platform drew to a close.
As with anything on the Amiga, I had little to no problems with it and I’m sure if they had been allowed to innovate at the level they were working at, we’d be lightyears ahead in computing right now.
Never owned a machine that rocked so hard as my Amiga. AWESOME machine, just AWESOME. Loved the C64, crazy about Macs, but the Amiga took the cake.
Right now, the best way to install apps is the Apple way. I don’t know how installers work for all platforms, but I’ve seen plenty of Windows installers and VISE. .pkg installing is the way to go for the Windows user. You just make your selections and when you’re done click the install button and you can have a cup of coffee [depending on the size of the app].
Clean installing, no hassle.
OR just dragging an app to the applications folder.
Anything you do more than one of these is a waste of effort.
The days of the Amiga. Guys, you’re making me nostalgic.
Those were the days of real hacking.
/mad props to Grovsnus and the crew
@ The flying boolaboola
You wrote:
>>
Right now, the best way to install apps is the Apple way. I don’t know how installers work for all platforms, but I’ve seen plenty of Windows installers and VISE. .pkg installing is the way to go for the Windows user. You just make your selections and when you’re done click the install button and you can have a cup of coffee [depending on the size of the app].
Clean installing, no hassle.
>>
It is the best way indeed, but the article wants to point the fact that users must have full control of what is happening into their machine environment else they were enslaved by their OS.
To avoid the problem of being taken unaware of the modification made into the OS, there is a solution.
I suggest (at least) that there should be an official OS “System Changelog” that every user could check, to verify if there were changes into system files, library or else during installation.
It is vital for the user to became quickly an experienced user and keep the control of the platform…
And this is more true in these days in which OSes are overwhelming the user himself and are notoriusly known to install all by themselves malicious programs which could even be considered Spyware, i.e. Windows Media Player (a program that transmits to an external server informations about your taste on music or videos when linked to internet)
Also on Macintosh you are unaware of bad behaviour by the manifacturer of the OS (Apple).
Do you know that on OS-X if you install latest free-version of Quicktime you lost the feature to run videos fullscreen, and that you are forced to buy a full version of Quicktime to view back in fullscreen mode???
Yes. This is not a real problem but only an example of what is the tendency nowadays in developing new System Software.
Because I know that (regarding Quicktime) immediately some people of the Apple-Macintosh scene released an open-source patch to avoid that lack of features of new version of Quicktime…
How lucky!
…BUT…
What about if the programmers of the patch were forced legally from Apple to drop the development of any program related to Quicktime?
You could be forced to buy new one or at least to re-install old versions of Quicktime.
So this is the problem!
It remains still open the question of achieve full control of installation procedures, to control what these “installs” are doing, and maybe EVEN to restore previous installed versions of software to avoid upgrades like that of Windows or even “stings” as that of QuickTime by Apple.
Easier is the OS, then faster any user could became experienced in it.
(AmigaOS is certainly invaluable about it)
>>
OR just dragging an app to the applications folder.
Anything you do more than one of these is a waste of effort.
>>
This is sure better than the first. The user must cope only with a single file.
But any new application must be written to fit perfectly the system you are using.
It is the programmer or the manufacturer firm which has the responsibility they must relase different applications (made of a single file) to fit different hardware solutions.
SingleApplicationFile-for-i386linux
SingleApplicationFile-for-AppleOSX
SingleApplicationFile-for-WindowsXXX
Regarding Windows it must also cope with different versions actually existing of the Windows OS (Win98, WinME, WIN2000, WinXP, etc.)
Even into Apple or AmigaOS environment the single-application-file shuold have different versions.
SingleApplicationFile-for-PPC-CPU-WITH-ALTIVEC
SingleApplicationFile-for-PPC-CPU-WITHOUT-ALTIVEC
And what if Apple will just tomorrow release new computers with minimal hardware differencies from the actual ones?
The difference could be the Video-card, or the Audio card. It is hard to foresee even little advancement in hardware.
Does the single-application-file will start again?
An universal application-file? This is a lucky event, and it not happens always.
It will be needed new versions of the software.
— Starting here, it follows my comments for AmigaOS and MorphOS users only —
I think that the better installation procedure described in these comments is that one in MSG: nr.1 by Mr. Hobgoblin who uses Gobolinux:
>>>
when it comes to installations i must say i have fallen in love with the system used in the gobolinux distro.
pop the new application into a version numberd folder under a folder named after the app. put a symbolic link next to the versioned folder, pointing to said folder and name the link current. presto its installed
>>>
That is:
user must create by himself the directory (or folder) and copy the files needed.
System disk -> homedir -> “new-application folder” -> “version folder” nested into first (for upgrades of the software)
Then the user needs only to create a symbolic link by himself to get the new application running.
If he wants to de-install it, then he only needs to remove symbolic link and delete new application directory.
The user has full control of any application!!! True he got it.
This Gobolinux distro sure it is raising my interest! I think this kind of installing procedure is a very decent system to increase ease of use of Linux OS.
Luckily and magically the Professor Peruggi (who wrote the article mainly for AmigaOS-MorphOS) has already realized a system to replicate on Amiga and MorphOS same kind of installing procedure shown for Gobolinux.
He donated the MorphOS community two AREXX utilities to Trap any Amiga installation procedure.
These utilities capture the new-application installed files and copy them all to another path or directory-tree which replicates original structure of Amiga-System Hard Disk.
I will try to explain it:
Imagine these two partitions scheme:
Amigasystemdisk:
C dir
Devs dir
Libs dir
Fonts dir
etcetera
Programs partition:
Newapplication directory/
C dir
Devs dir
Libs dir
Fonts dir
etcetera
With these AREXX scripts, the user could trap new installed programs into the doppelganger directory tree…
Then the user, he just needs to type correct “ASSIGN command string” into S:User-Starup script
(for those who are not acquainted with Amiga, the User-Startup is a script that AmigaOS uses to load at boot variables, preferences or any program chosen by the user)
Assign programs:new-application-dir/Libs Libs: add
Finally the program will magically start as if its files and libraries were laid into system directories.
(The “Assign” command in AmigaOS is used to link new libraries or directories to the search path of standard system ones.
It creates sort of symbolic link similar as those in Unix-Linux OSes.
Then “newapplication-dir/Libs” will be virtually added to the search path of Amiga System-Libs called “Libs:”.)
>Anyway, note that this discussion is absolutely boring for
most people here (I’m sorry for that). In fact, why should
they care for MorphOS and our negligible niche?
What’s sure is that no one will care as long as this kind of discussion goes on… But this doesn’t mean someone will care as soon as these talks are ended… Amiga/MorphOS is today years behind any Linux, Win, OSX…
Leo.
Excuse me.
Assign programs:new-application-dir/Libs Libs: add
should be
Assign Libs: programs:new-application-dir/Libs add
>>>
Anyway, note that this discussion is absolutely boring for
most people here (I’m sorry for that). In fact, why should
they care for MorphOS and our negligible niche?
>>>
>
What’s sure is that no one will care as long as this kind of discussion goes on… But this doesn’t mean someone will care as soon as these talks are ended… Amiga/MorphOS is today years behind any Linux, Win, OSX…
>
It is a matter of fact that almost all readers here on OSNews live in free countries and I am sure that nobody forced you to read news regarding other operating system you don’t care of.
And I find very curious that you were so “positively” bored of this thread, that you even need to comment it.
😀
Well, sure I must admit the fact that to deal with installation procedures on every OS actually on the market is a topic really booooooooooing.
> Their interest may be attracted only by Pegasos computers,
> an open hardware that may free them from x86 dependency.
Could you please explain what Dependencies you’re talking about?
Further more, what ist an “open hardware”?
thx.
> …x86 dependency…
> Could you please explain what Dependencies you’re
> talking about?
People that do not want to support the Microsoft/Intel/AMD
triumvirate may be happy to use other OSs and processors.
This is not the old stupid invitation to leave a platform
because “it is bad” (whatever this can mean in technical,
economical, industrial, or political terms). Everyone is
free and will use what he judge the most convenient product.
Anyway, it is my opinion that people should always have at
least two possible choices in any field. Beware of the
monopolies! People without choices are already slave, even
if their masters are very nice persons.
So, I appreciate very much the fact that people who don’t
want to use Windows on x86 hardware can use Linux on that
hardware. And I think it’s highly desirable that people that
don’t want to use PCs powered by x86 and related processors,
whatever are their reasons to do so, can use PC’s powered
by, e.g., PowerPC processors.
> …open hardware…
There two alternative choices in the PPC processor area:
MacIntosh computers and Pegasos computers. The former are
much more powerful and up-to-date (also more expensive) than
Pegasos computers. On the other hand almost every
information about Pegasos hardware is available; this is not
true for Apple hardware.
No judgement is implied; no implicit or explicit
interpretation is provided. I only report some facts and
also do not pretend that others agree about their relevance.
Again, what’s actually important for me is only that, even
in this case, we are lucky because there are two alternative
choices.
Sorry to be so picky, so long, and so neutral. But I don’t
want to be involved into endless and useless “philosophical”
discussions.
Cheers.