Gentoo, Fink, and DarwinPorts Join Forces

Mr. Quick writes "From Metapkg, "In order to better provide freely-available software to users of Mac OS X and Darwin, we Fink, Gentoo, and DarwinPorts commit ourselves to work together." A unified front for free software on Mac OS X is something that was needed."

Comments

Uh.. so

[Anonymous Coward]

So while this is really cool, how is it going to work out?

To wit: thought maybe i'm on crack, it SEEMS like each of the three-- while offering basically the same interface to the same service-- were pegged to different codebases, and taking packages from different sources. Fink to debian, gentoo to gentoo and ports to bsd.

Is this the case? And which source (debian/gentoo/bsd) will the collaboration generally follow?

This is what Linux needs

[Anonymous Coward]

...to come together like this. The competing GUI's (KDE vs. GNOME), the competing browswers (Konqueror, Mozilla, Opera, Galeon), the competing distributions (SuSe, RH, Caldera), all drain human and financial resources that, if combined would make Linux into the powerhouse it could be.

Until then, Linux will remain second fiddle to the likes of Windows XP and MacOS X.

Re:This is what Linux needs

[HornyBastard77]

Yes indeed. Get them all together. Package them all the same. Get rid of choice, it is overrated anyway. That is the only way to get Linux to be just like Windows, and the OSI, FSF et al to be like just like MS.

Re:This is what Linux needs

[90XDoubleSide]

These group's aren't merging into one project: they're still making three products, but will be working together to ensure there is sharing of work and no duplication of effort. This is indeed different from most OSS projects where the two competitiors come to hate each other for some reason. Another groundbreaking thought from their mission statement: "Non-advocacy: Our common goal is simply to provide software for people who choose to use Mac OS X & Darwin, not to promote or advocate any particular operating system." OSS with non-advocacy! Imagine how much more acceptance open-source software might get if everyone focused on telling people how the development model could produce great software at no cost to the user instead of droning on about how it is immoral for programmers to serve as wage-labor.

Re:This is what Linux needs

[Idimmu Xul]

What choices would you be losing if that happened? They same software would still be available, and you would still beable to do the same thing with it, e.g. install it from packages, install from source packages, roll your own from a tar ball.

The only thing that I can see that you'd lose is the multitude of different ways things can be configured. e.g. is the httpd.conf in /etc/apache /etc/httpd /var/www/conf or somewhere else? etc etc

Where is the choice between GNOME and KDE when you have to have both installed anyway to beable to use all the decent apps avilable to Linux?

Re:This is what Linux needs

[SilentMajority]

There are pros & cons to this.

The drawbacks to having fragmented marketshare (like KDE & GNOME) is sometimes--but not always--outweighed by the improvements caused by having strong competition.

Look at the drastic improvements MS IE received while Netscape was still a strong contender. Then look at the improvements after IE got 90%+ marketshare. Some would argue that there isn't much to add to a browser but a look at the innovations in Opera, Mozilla Firebird and Safari.

Rather than consolidation, I'd rather see competing products like KDE & Gnome come up with common standards. For example, KDE & Gnome could come up with very specific & consistent user interface standards and adhere to them in their products.

Microsoft did a great job (compared to Linux) in not only coming up with Windows UI standards but in preaching it: the vast majority of Windows apps writting by diverse vendors has a FILE, EDIT, HELP, etc. menu and they are rather consistent in their content too. CONSISTENCY IS IMPORTANT.

I'd like to see Linux be different where it counts: like stabiliy, security, open standards, Unix-like shell & filesystem, etc. But I don't see the point of being different for its own sake (like throwing out MS Windows GUI/UI guidelines so that 95% of pc users will find it less desirable to switch to Linux).

My apologies if such a GUI/UI standard exists--I simply don't see it being promoted or used in X apps I've tried--and it was just an example.

Re:This is what Linux needs

[Telex4]

Why do these trolls keep getting moderated up to 5 points? Let me go over this again:

o Duplication of effort, whilst often unecessary, is not the death of Free Software. Maybe some people don't like the code architecture of KDE, and so choose to code for Gtk and GNOME; maybe some people dislike the lack of challenge in approaching Konqueror and so choose to code on Mozilla. It's people's choice, and I see no reason why we should try to force all FS developers to code in particular projects.

o Many projects with the same goal is also not that big a problem. Look at the many different ideas that KDE and GNOME have developed, and then shared when good. Relations between the projects are improving a lot, to the extent that now KDM and GDM (the login managers) are sharing certain configuration files!

o Different distributions are also always a good thing. They each provide a different focus, with different strengths and weaknesses. Whilst this might be more confusin than one GNU/Linux distribution, consumer choice is an important thing. When you buy a car, you research the many options and chooce the one best for you. Why should we have to put up with such a limited choice in operating systems?

o I could go on but my food is ready ;-)

Re:This is what Linux needs

[Japer Lamar Crabb]

Except that this is merely pooling efforts on the porting side, so that one's choice of package management tool won't restrict availability of a given piece of software anymore than one's choice to use PPC Darwin.

Re:This is what Linux needs

[hhw]

My thoughts exactly... I think this is one of the reasons the 3 BSD's are able to accomplish so much despite having a fraction of the developer resources of the Linux community. Although each project is developed independently according to their respective goals, they share their solutions that the others are free to adopt or improve upon.

Re:This is what Linux needs

[Sentry21]

The drawbacks to having fragmented marketshare (like KDE & GNOME) is sometimes--but not always--outweighed by the improvements caused by having strong competition.

The problem is, as you point out later, inconsistancy. You can't troubleshoot Linux nearly as easily, because you have to say 'Ok, go to the menu with the foot on it, and choose- what? Oh, ok, the menu with the K on it. And then go to, uhh.. let me see, I only ever use Gnome.' Having to know twice as much can interfere ver much with helping someone.

Rather than consolidation, I'd rather see competing products like KDE & Gnome come up with common standards. For example, KDE & Gnome could come up with very specific & consistent user interface standards and adhere to them in their products.

As I recall, they agreed to work together on this a while ago. I could be wrong, but it's in the slashdot archives I'm sure.

Microsoft did a great job (compared to Linux) in not only coming up with Windows UI standards but in preaching it: the vast majority of Windows apps writting by diverse vendors has a FILE, EDIT, HELP, etc. menu and they are rather consistent in their content too. CONSISTENCY IS IMPORTANT.

Microsoft did a horrible job with their GUI. The standards you mention (File, edit, help) have been around since the early Mac days, and are in a slew of DOS programs too. Microsoft, however, made a lot of very bad design decisions - such as using 'Yes/No/Cancel' or 'OK/Cancel' dialogs whenever a choice needs to be made, instead of properly labelling the buttons with exactly what they do.

They also don't stress the importance of making one's program follow the same pattern as the 'standard'. Most programs, when you try to close them, have a 'Save? [Yes/No/Cancel]' dialog, but enough of them have an 'Abandon changes? [Yes/No/Cancel]' dialog to make life frustrating for anyone who deals with a wide variety of programs. If you want real UI guidelines, check out the latest ones from Apple. It's a near-religious text.

I'd like to see Linux be different where it counts: like stabiliy, security, open standards, Unix-like shell & filesystem, etc. But I don't see the point of being different for its own sake (like throwing out MS Windows GUI/UI guidelines so that 95% of pc users will find it less desirable to switch to Linux).

Linux environment programmers (KDE, GNOME, etc) have three main options. If they copy the Windows behaviour, it'll be familiar to Windows users, but the Windows behaviour makes little sense in a lot of circumstances (See above)

If they (properly) copy the MacOS behaviour, they will have a system that feels and works properly even to a completely new user, but most people consider it 'wrong' because it's not what they're used to. As such, it will probably never be adopted, since a lot of people refuse to give it a chance.

If they make their own guidelines, then obviously, they'll have the freedom to make their own standards, which they can tailor to suit their programs. This is bad, but it's what will happen. As evidenced by Sun's usability study, programmers design interfaces for themselves and others, but don't tend to consider what other people are used to or will find intuitive - well, how would they know? It makes sense to the programmers. A lot of programmers consider the UI an 'interface to the user for the code' - a way for the code to get itself run - rather than 'an interface for the user to the code' - a way for the user to make the code do what they want.

What the GNOME/KDE projects need are clear heirarchies, and priorities. Unfortunately, GTK is such a rabid bitch to code in (compare to Cocoa) and not many programs use Glade, so UI designers have the harsh end of the stick. I also can't help but feel there's a feeling with GNOME programmers that only 'real programming' is beneficial - documentation and UI design can be 'good enough' (docs and UI can never be 'good enough'). If it's there i

What about Apple?

[BibelBiber]

Since they ported X11 to Mac OS X on their own it would be kinda useful to have them in the same boat. Dont you think?

Re:What about Apple?

[majorflaw]

Apple's #1 priority is selling Apple hardware. While they make some $ through the sale of software, harware is far and away their major source of income. I would expect that Apple is delighted when someone else writes software that works well on the Mac platform; just another reason for people to buy a Mac. The difference between Apple and their carnivorous competitor is that Apple doesn't really care what you run on their computer, as long as you buy the computer.

Coordination in Open Source development.

[dwerg]

I think most people don't understand how unique this initiative is. Most of the times open source projects don't really notice eachother and when they do, they just start a flamewar about who's best and who stole feature from who.

It's good too see there are some developers out there with organizational talents who are willing to communicate with other projects in order to speed up development time and create a better product.

Re:Coordination in Open Source development.

[ndogg]

I think that the flamewars come more from non-developer users of OSS software rather than from the developers themselves. Take, for example, KDE and GNOME; it's the users that bicker, not the developers. In fact the developers want to work together and port features across. There are some arguments that happen across the camps, but that usually comes from differing philosophies about the UI or just conflicting personalities (which happens with any organization.)

So, I guess the real question is: why do end users (i.e. people who just use the software and do no development on it) bicker so much? I'm not quite sure, really.

Why not just port GTK/Gnome to MacOS X ?

[Anonymous Coward]

That why you just compile your favorite GTK/Gnome app and have a native MacOSX app ?

Re:Aren't they forgetting someone?

[MalleusEBHC]

...but running standard unix apps under X on top of OSX doesn't take advantage of OSX's strong points.

Unless I'm alone here, being able to run X11 apps and native OS X apps at the same time is one of the best features of my OS X boxen. The availability of diverse software from two almost totally separate camps is awesome.

Excellent news

[harikiri]

One of the reasons I've installed Yellow Dog Linux on my iBook (for coding & development) was because it was such a pain having to search across multiple "vendors" of open source ports and packages for Darwin. Depending on which package I installed, I would either have to modify makefiles to use up to three different -L (path's to programming libraries), such as /usr/lib, /sw/lib and /usr/local/lib. It was bloody annoying.

So I welcome this move towards a unified ports system for Darwin, it was definitely needed.

Re:Gentoo Translate-O-Matic

[Doobian Coedifier]

It's sad that Slashdot has gotten bored bashing windows, now the distro wars are heating up.

It's even sadder that you post this crap that has been posted verbatim several times before (that I've seen), and you didn't write.

And yes, I use Gentoo. And yes, it DOES kick ass.

Why can't they use BSD's system

[Billly Gates]

kiss= keep it simple stupid.

There is no need to make a complex metapackage system.

I find Gentoo's python based system way overly complex and buggy. You need to emerge rsync quite a few times during a new install to ensure you are using the latest version of portage.

The FreeBSD ports system on the other hand are just simple tcsh scripts. Under /etc/defaults/make.conf you can specify which mirrors to use for popular ports or you can type in the closest FreeBSD ftp site and over-ride it for the fastest download speed.

If any of you reading this use FreeBSD 5.x go to /usr/local/examples/etc/defaults/make.conf and edit, cut and paste the data to /etc/defaults/make.conf. For some dumb reason the FreeBSD team moved all the import rc scripts there. The big commented scripts is one of the traditional strength's of FreeBSD. I hate it when they make it harder for newbies.Do a man make.conf for more info.

WHen you do a "make install clean" the port scripts just use standard ftp and http sites in the makefile to download the apps. Nothing complex and its alot easier to use.

I can not speak of fink because I have never used it.

Simple shell scripting can get rid of alot of complexity.

Re:Why can't they use BSD's system

[Fros1y]

Well, so far as I can tell, the complexity of the Gentoo system is well-bounded, with nice python scripts not all that much more complicated than your average shell script (Assuming you're able to program, of course.

And I can't really imagine what you're talking about with "You need to emerge rsync quite a few times during a new install to ensure you are using the latest version of portage." Granted, installation has quite a few steps (For the most customized, only) but I don't remember ever having to 'emerge rsync quite a few times' in any of my multiple gentoo installations.

As far as the complication of day to day use, I can specify standard mirrors for my gentoo system to use too. In fact, using mirrorselect, I can do it automatically to pick good mirrors for me. Then again, I don't even see how 'make install clean' is so much simpler than 'emerge foo'.

Frankly, I'd rather wrangle python scripts than shell scripts anyday. I find the idea that shell scripts eschew complexity kind of off. Gentoo has been able to grow technically so well, I think, in part of the advantages of a full and elegant language (Python) of implementation, rather than a hodgepodge of shell scripts.

Just my two cents as a Gentoo user saving up to be a Gentoo OSX user.

Re: rpm + apt-get

[autechre]

Or you could have a distribution which has modular packages instead:

apt-get install mozilla-browser

I don't have to recompile PHP every time I want to use a different module; I just install whatever modules I want, whenever I want to use them.

That is the strength of Debian. It's not just apt-get; people who have ported apt to work under Red Hat are moving in the right direction, but that is not the whole problem. With Debian, thousands of packages are "official", and so are quite strictly designed so that all dependencies really, really work. The organization of packages is what really makes apt worthwhile.