Why isn't Arch more popular

scottro · 2003-01-12 22:47:24

Well, if I understand your point, (that is, if you're referring to developers seeing postings) I suspect that it will have a natural evolution.

As more people join, and posts increase, user X will see user Y's post, and have the answer. In other words, it won't be as necessary that the head developers see the posts.

Also, as a user base increases, you'll get more and more newcomers--there will be more and more
questions that have already been answered 20 times, or that are simply answered with a little
research. That's the nature of the beast.

BTW is anyone on this list going to Linux Expo in NYC?

Scott

dunbar · 2003-01-13 14:19:02

Cooperative help is where Arch can grow.

No need for me or everyone to become one of the haxor l337 (as if).

As a forum member, whenever I see the same problems posted repeatedly, I can either answer it with a 'go thither' message (pointing to the thread with the answer), or, if issues are not yet resolved, I have some access to the top dawgs, which access is based on my demonstrated credibility.
Example: If I'm reading a post from wiggienewbie, this is wiggienewbie post #2 and the post is way off base, wiggienewbie has low demonstrated and low credibility, based on my familiarity with wiggienewbie, the subject matter and their quantity of posts. OTOH, if it is ody posting an answer to a question, he gets top dawg status for both demonstrated and credibility (subject matter determines whether I'll even read the post).

I know ody, I don't know wiggienewbie, so I might ignore wiggienewbie.
Standard forum ranking stuff, occurs everywhere.

The determination of who gets answered and who gets ignored is made by the reader.

(Heirarchy - thats the word.)

Same applies to PM: if I PM ody, reader ody will decide whether and how to answer based on the specifics of the matter, which specifics may include whether I have a demonstrated credibility problem.

Yes, in the future, a larger user base creates larger quantity of posts and a larger quantity of PMs, but we tend to rank these and delegate responses to the lower echelon, as scottro indicates.

Have no fear: we will grow and ranks will develop.

mordantly · 2003-01-14 04:19:35

hax0ring is bad. mmmk. dont hax0r. mmmk. that is a quite valid point... accept than the new kid in the crew doesnt have a fair (or stand a) chance nessecarily to speak and seek an answer do they?? i mean if u (joe blow member) is scannin the posts than u may or will skip that post as it will have been predomanently filed under 'bullshite' no??

sarah31 · 2003-01-14 05:47:55

actually technically hackers are not bad. it is crackers that are bad. hackers generally do what they do for interest and learning sake. they will not "attack" a "joe user" . nor are they out to wreak havok on the world.

crackers, on the other hand, are a cancer that needs to be cut out.

beniro · 2003-01-14 23:45:33

Hello. I think I'm dunbar's "wiggienewbie".

The best way to make Arch more popular would be to nurture an Arch Community as much as possible, I believe. Display prominently all of the energy and optimism surrounding the project. Possibly make an attempt to keep the front page a bit more updated, even if that means posting more minor news, like: "wrestling with a particularly tough problem in blah today. i hope to have this ready by bleh." People will see more recent news items and the project will seem to have more vitality.

Or whenever there is a new release, link directly from the news item on the front page to a forum thread discussing that release. For instance, to the end of the news item on the release of .4, you could just tack on "Discuss this here". It would help get more people talking on the forums.

Also, I think the front page could maybe be redesigned in such a way that it would "sell" Arch a bit better. Maybe have a "Why Arch?" link or something to that effect. Make that logo not necessarily bigger, but situate it in such a way that it calls the attention more effectively, because your logo gives a potential user something to grab onto, something immediate to identify with. After all, even us nerds like purty pictures.

Of course, I realize this takes even more man hours...

Second most important issue for increasing Arch's popularity is to work on the documentation. This will naturally start to come along as Arch nears version 1.0, I believe/think/hope.

All this being said, I think Arch will grow naturally, because quality products are always in demand...

But there's no harm throwing a bit of fertilizer on the roots of this little community.

sarah31 · 2003-01-15 00:25:30

yes very good points. points that i have not considered myself. i don't know what it would require man hours-wise but seems to me to be worthy of consideration.

xirus · 2003-03-01 12:18:02

to answer the initial question, besides maybe some polishing (jk's posts) I think the only thing that can possible keep people away from this distro is the lack of applications

this will be solved when arch becomes more and more popular ->
user x needs apps a, b, c and will ask for them or build them himself
user y needs apps a, d, e and will ask for them or build them himself
user z needs apps b, e, g and will ask for them or build them himself
this way more and more apps will be submitted (and we will need more package maintainers)

coming from gentoo, arch has a better installer, a better package system (that is: binary and not source based - I think pacman needs a lot of improvements but there's being worked on that), is easier to configure,...
There's no reason to stay with gentoo except for the apps

sarah31 · 2003-03-01 14:51:29

=X¥®µ§= wrote:

I think pacman needs a lot of improvements but there's being worked on that), is easier to configure,...
There's no reason to stay with gentoo except for the apps

just like to comment on these:

yes pacman needs some polishing but since you mentioned gentoo, i would say it is miles better than gentoo's emerge. though emerge works it still has security risks in it that make it quite easy accidenttaly foo bar your emerge. i always hear of it breaking. unmerging is a huge risk in gentoo much more so than pacman -R is.

okay gento has more apps but it is also an older distro than arch. it also has a horrendenously managed packagae tree. i understand the reason for masking, for example, but why even offer the package if you KNOW it is not stable? if only certain people should use a package why offer it? i gave up on gentoo not because they had a small package tree but because that package tree was horrible once i had my system in i felt cripple because alot of what installed was very unstable or down right broken. the other 50% was masked. it was a nightmare for two whole weeks.

finally not using any distro because it does not have what you need package wise is a bad excuse. with arch for example it is easy to make 90% of what you want. sadly only five or six peopl regularily make packages and they will only make what they need and as a result it does not mean as wide a set of packages as some may want, but arch allows the user to actively contribute not many distros will permit user contributions.

more maintainers. with what is in official, unofficial, and soon unstable we currently have enough maintainers. i stay ery busy with main taining and fair ly well ontop of package verions. I have not had much of a chance to add packages recently or build but those are only a small part of my job here.

xirus · 2003-03-01 16:43:26

Ok, I believe you when you say that pacman is better under the hood!
And you are correct, my gentoo system was highly unstable in the end, it crashed more than windows 98 did a couple years ago...

However I'm talking as an end user:
1) package groups so you don't have to install all kde apps manually but you just say: pacman -S kde and it installs them all (but I know that there is being worked on this or it's maybe already done)

2) a hierarchy of programs, like other distros do it
http://www.gentoo.org/dyn/pkgs/index.xml if you are searching for eg a webbrowser, you just go to net-www and there you have a lot of options. Or http://www.freebsd.org/ports/index.html or http://packages.debian.org/stable/ and so on.
ATM archlinux's system will do, because there are so little packages, but once you get a lot of them (and I'm truly confident that this will be true within a year) you can't see the trees through the wood anymore (to say it in a belgian expression). It's a lot cleaner to group packages!

3) I would like to see if a program belongs to unstable, current, stable or testing when I do a -Ss -> if it's in unstable I might not install it at all
And if you install program X wich has dependencies (wich are showed when you are asked if you are sure you will install the application) I would also like to see where these dependencies belong to

4) always having to use 2 chars instead of 1 as pacman options is not really user friendly and confusing to beginners... why not just use "pacman -s" to search for a package??

5) it takes very long for pacman to search for packages - I think this is caused by using the folder structure as opposed to putting all programs+description+dependencies in 1 file?

6) there were some other small things but I can't remember them atm

These are just my 2 cents
Don't misunderstand me sarah, I will continue to use arch instead of gentoo (because of the compile time, the unstability and all the advantages of archlinux) because I really believe it is better

However as you already said: arch is still very young so it IS normal that there are things that need to be polished. Every person is different so each user has his/her own complains or wishlist. The 5 points above are mine regarding pacman, what happens with them is up to the developers of course

<i>(I might not answer to a potential reply (that might correct me on some points) -> this is not because I'm rude but because I'm away for 10 days, so don't be offended if I don't answer your comments immediately )</i>

sarah31 · 2003-03-02 06:36:55

3) I would like to see if a program belongs to unstable, current, stable or testing when I do a -Ss -> if it's in unstable I might not install it at all
And if you install program X wich has dependencies (wich are showed when you are asked if you are sure you will install the application) I would also like to see where these dependencies belong to
something like this you mean :
sarah@ArchLinux:~/Arch/unofficial/network/wlan-ng> pacman -Ss mysql
current/mysql 3.23.55-2
    A fast SQL database server
unofficial/mysql4 4.0.10-1
    A fast SQL database server

sarah@ArchLinux:~/Arch/unofficial/network/wlan-ng> pacman -Ss mozilla
current/flashplugin 6.0r69-1
    Macromedia flash plugin for Netscape/Mozilla (nonfree)
current/mozilla 1.2.1-1
    The Mozilla Web Browser
you should notice that with the newest -pacman the search option is quicker---much quicker possibly more so than debians search. as for dependencies here are the rules conveyed to me by apeiro:
official/current packages will NOT have dependencies outside the official/current tree (ie if you install a ANY package from the official tree all of its dependencies will be from official too.
unofficial can have dependencies in official/current or unofficial but not in unstable.
unstable can have dependencies in all the trees.
2) a hierarchy of programs, like other distros do it
http://www.gentoo.org/dyn/pkgs/index.xml if you are searching for eg a webbrowser, you just go to net-www and there you have a lot of options. Or http://www.freebsd.org/ports/index.html or http://packages.debian.org/stable/ and so on.
ATM archlinux's system will do, because there are so little packages, but once you get a lot of them (and I'm truly confident that this will be true within a year) you can't see the trees through the wood anymore (to say it in a belgian expression). It's a lot cleaner to group packages!
my own personal preference is not to have too many subroutines for a package search. while i understand it and appreciate such heirarchical organization i found that too many subdivisions is confusing when i was doing searches for debs back in the day i found that it was not so easy to go to foo directory in the heirarchy and be absolutely sure to find a specific app i was looking for. i prefer that such searches be saved for web site and arch already has a heirarchy that if ind is diveded up into catagories that are logical. they may be large in the end but having www-browsers, www-net, www-foo, www-bar, and so forth is getting just a little over the top. debian has less catagories in their hierarchy but still it can be hard to find what you are looking for.
i dunno it is a hard area to get setup so that the majority are pleased. while it is nice to have such features it I would not like such features to make the package manager slower or kludged up with options.
1) package groups so you don't have to install all kde apps manually but you just say: pacman -S kde and it installs them all (but I know that there is being worked on this or it's maybe already done)
task/meta/group packages are something that is a must i think BUT i disagree the ALL kde or gnome (for example) packages be installed. i would think that only those packages needed to run and manage the grouped package be installed and the "fluff" such as games or extraneous tools, etc be left as single, optional packages or their own separate groups. of course the option to install group packages ungrouped should remain as an option.
4) always having to use 2 chars instead of 1 as pacman options is not really user friendly and confusing to beginners... why not just use "pacman -s" to search for a package??
i agree but at the same time i don't think it is possible to have single options only. it is nice to have a small simple set of all lower case options but what happens when you want to do several options at once? well one option is to start having single letters perform what several letters do. that means one could potentially have 26 options for pacman. not entirely easy to remember but entirely possible that as pacman evovles to have this happen. what then happens if you need more than 26 options? well start using capitals i suppose. ....
i assume you see what i am getting at. it is nice to have single options but at the same time it is so much easier to remember eight options and just combine them in series, like -Syu.
5) it takes very long for pacman to search for packages - I think this is caused by using the folder structure as opposed to putting all programs+description+dependencies in 1 file?
as mentioned earlier you should find the newest pacman much quicker at searches. as for your reasoning why pacman was slow at this function, i don't know absolutely, but i don't think that the folder arrangement is the culprit. think apeiro mentioned what was the culprit in a previous thread and i don't think the folder arrangement was what he mentioned.
the above answers are my opinions and by no means take them to mean that i don't think there is no room for improvement in pacman. i think there are things that can be done but they should be done with due consideration. i would hate to see an option added that ends out stunting other established features. i would hope that even while the package tree grows that pacman will stay small simple and fast.
many of the features you mention would be very nice to have and i hope to see them added and pacman become the very best regarded package manger in linux. (which i think it is well on the way to becoming)
AKA uknowme
I am not your friend

Offline

Arch Linux

#26 2003-01-12 22:47:24

Re: Why isn't Arch more popular

#27 2003-01-13 14:19:02

Re: Why isn't Arch more popular

#28 2003-01-14 04:19:35

Re: Why isn't Arch more popular

#29 2003-01-14 05:47:55

Re: Why isn't Arch more popular

#30 2003-01-14 23:45:33

Re: Why isn't Arch more popular

#31 2003-01-15 00:25:30

Re: Why isn't Arch more popular

#32 2003-03-01 12:18:02

Re: Why isn't Arch more popular

#33 2003-03-01 14:51:29

Re: Why isn't Arch more popular

#34 2003-03-01 16:43:26

Re: Why isn't Arch more popular

#35 2003-03-02 06:36:55

Re: Why isn't Arch more popular

#36 2003-03-04 12:43:15

Re: Why isn't Arch more popular

#37 2003-03-05 12:17:25

Re: Why isn't Arch more popular

#38 2003-03-09 20:23:46

Re: Why isn't Arch more popular

#39 2003-03-09 20:35:48

Re: Why isn't Arch more popular

#40 2003-03-09 21:33:59

Re: Why isn't Arch more popular

#41 2003-03-09 21:47:59

Re: Why isn't Arch more popular

#42 2003-09-12 02:58:41

Re: Why isn't Arch more popular

#43 2003-09-12 04:34:03

Re: Why isn't Arch more popular

#44 2003-09-12 07:16:03

Re: Why isn't Arch more popular

#45 2003-09-12 17:44:58

Re: Why isn't Arch more popular

#46 2003-09-13 04:58:28

Re: Why isn't Arch more popular

#47 2003-09-13 05:29:18

Re: Why isn't Arch more popular

Board footer