CrunchBang Linux

omns, rhowaldt, sqlpython - thanks, it is nice to feel welcome

I admit there are possible downsides with the particular method I chose to use during discourse such as this, such as missing the forest for the trees so to speak.  So long as one makes a conscious effort to avoid this, though, I do not see it as a serious issue.  Moreover, quoting exactly the items to which I am referring seems to significantly lessen the frequency of misunderstandings.

This issue with this is that you never explained the logic behind any of you conclusions, only the conclusions themselves.  Given such a situation, I have no way of differentiating between two possible interpretations of what someone such as yourself could be saying: (1) You are changing your mind over the course of the discussion (as now seems to be the case), or (2) that you are presenting a single message and, from the fact that I have seen what appears to be conflicting opinions, I have been unable to correctly decipher what the message is.  Assuming the second case seems to always be a safe option.

I can certainly understand time constraints.  Finding the proper balance of clarity-of-message versus time-investment is a difficult technique to master.  Personally, I believe you are swaying to far onto saving time at the cost of clarity, but know that I am aware of your choice on the matter I will be better suited to interpret what you say accordingly.

I almost always read the entirety of a message before responding, and I place a conscious effort on attempting to see the over-arching message.  I seem to have failed to understand you despite that fact, and I disagree that the way I quote individual concepts to which I am responding was at fault.  The way I have quoted individual sections was intended to highlight exactly where I got lost in what you were saying.  Whether I failed to get your message because you were not clear, reading comprehension issues on my part, or a combination of the two is certainly up for discussion.

I am certainly quite glad to hear that, as I meant absolutely no offense what so ever.

I do not follow why they would be upset, nor do I see how I created questionable content.  I failed to understand what was said, recognized this, and explained in as clear of terms as I could where I got lost to ease potential efforts to remedy the confusion.  I am unaware of any superior way of going about such discussions.  Dropping the quoted sections would only serve to increase potential for further confusion - I am completely missing how it would help at all, beyond possibly saving time.

They are often related, but not necessarily dependent.  It is quite possible to have a secure system which the user knows very little about, although this usually comes at the price of functionality.  See, for example, Apple's attempts with their various touch-screen-based products.  While those do have security issues, the issues are not necessarily a result of the end-users not understanding them.

However, that is tangential to the actual discussion at hand.  The way I responded to each part separately in no way reflected that I thought they were completely unrelated - it was just I felt it best to make it extremely clear which aspect of the expertise-security-complexity system to which I was referring.

No offense was taken.  If I may be blunt, I have no idea who you are, nor do I have any reason to give weight to your opinion - you could loudly proclaim your contempt for me personally and I would not take it to heart.  I simply wanted (1) to learn why you felt the way you claimed to, on the off chance you did in fact have good reason to dislike Bedrock Linux, such that I could potentially fix the issues, and (2) to ensure what at one time was a highly-ranked result for "Bedrock Linux" which spoke ill of it did not falsely represent my Linux distribution.

I recognize the problems that often come paired with complexity.  Keeping the system as simple as I can while retaining the desired functionality is a very important goal for Bedrock Linux, as I attempted to state here: http://bedrocklinux.org/introduction.html#Simplicity

Bedrock Linux provides a number of benefits in exchange for the apparent complexity.  In my experience, the more complicated filesystem layout results in significantly simpler maintenance, despite the fact that it is still an alpha with a number of known issues.  For me, that results in a net gain, and will be doubly so once the issues are ironed out.  I certainly recognize that will not be the case for everyone, however I find it hard to believe I am the only person who feels this way.

I really hate to make this personal, but if what Bedrock Linux does is that complicated in your eyes, I call into question your ability to manage Linux systems "ranging from SOHO to up to small warehouses full of servers."  I admit it is weird, but it is entirely based on techniques of which any competent Linux sysadmin should already be aware.  Ultimately, for many system administration tasks, it should be significantly easier than many other distributions.  Not all, mind you, but many.  Note that this is assuming I work out all of the bugs that are keeping this in alpha.  This is roughly equivalent to being scared of a distro with a large repository - no one could possibly be familiar with every package in, say, the Debian repository.  However, one doesn't need to be - just don't install things you aren't familiar with.  Bedrock Linux with, say, two clients, both Debian stable, is fine.  You don't have to install every distro under the sun.  The vast majority of the time I am using three clients.

Which makes sense, since you seem to dislike it.  Personally I'd love to see it take off.

I suspect you are exaggerating, but if not, you may want to seriously consider such a route right now, as techniques such as virtual machines are already in heavy use and I don't see computers getting any simpler as time goes on.

Yes, please do so - that is what I really wanted from the beginning.  Simply stating whether you do or do not like something provides me very little value if it is not paired with the rationale behind the statement.

Your issue here, as I understand it, is that Bedrock Linux, if successful, will lessen the minimum needed skill-set to use Linux-based systems.  This would, in turn, would lessen the number of people who feel obligated to learn how Linux works under the hood, and, consequently, it would shrink the population which is able to contribute back to the Linux community.

Honestly, that is a good point I have not considered.  Touché.  I'll have to put some thought into this.  This is exactly why I do not immediately discount what may at first appear as uninformed, baseless hate, but try to give people a chance to back up what they are saying.  It is quite possible there is a solid point buried somewhere in there, even if it usually turns out to be nothing more than rash assumptions resulting from people not fully researching the topic or trolling.

I suspect perhaps you are simply reflecting your own opinions onto me; either that, or I am completely missing how you have come to this conclusion.  The Linux ecosystem is awesome.  It is not an entirely rare event that I sit back and reflect on just how much amazing software is available to me, free - with both definitions of "free".  I would gladly have continued to use it if I never thought of Bedrock Linux.  However, the Linux ecosystem was never perfect - there is room for growth.  What I consider the ideal Linux distribution requires far more manpower than I alone can provide or muster.  However,  I found a way to get something quite close which I could create by myself, and figured others would be interested in it.  Honestly, Bedrock Linux would be absolutely useless if the Linux ecosystem sucked.  Bedrock Linux is amazing because it stands on the shoulders of every other Linux distribution out there.

Quite reasonable.

Restating what I said above: if you aren't simply reflecting your own opinions onto me, I have no idea how you came to this conclusion.  Moreover, if you are reflecting your own opinions onto me, I strongly disagree with you.  The quality of many of the existing distro maintainers is so awesome that I readily bow to their work before my own at every opportunity: http://bedrocklinux.org/introduction.ht … 20Features

I will agree that this mentality is unhealthy.  However, those Windows users do it because, given their knowledge and time, this is the best way they know to defend themselves and yet remain functional members of our technology-based society.  It is more something to be pitied than to consider dangerous, as I see it.

This is the section which resulted in me earlier suspecting you are reflecting your own opinions onto me.  I completely disagree - the Linux ecosystem is incredibly healthy and strong.  In fact the best single word I would use to describe it is "resilient".  If, hypothetically, all of the maintainers for a given distro - even a major one - are paid off by Microsoft or go crazy or otherwise stop working productively on Linux - it other distros will remain to carry on.

I do not want them to stop, not at all - this is extremely far from anything remotely resembling a protest.  I honestly considered it more of a monument to how much respect I have for all of the other Linux distribution maintainers out there than what you describe.

While I do not feel this in any way describes how I feel about the Linux community, I can certainly understand the sentiment when applied to other things.

Honestly I feel Bedrock Linux has room to bring something of value to the community because of the conservatism you stated earlier.  Linux distributions function the way they do because (A) it works, and (B) people are very hesitant to try something new.  There seems to be an assumption that things work the way they do because people before us knew better than we do - best we not stray from their plans lest we stumble upon some issue they have foreseen.  Happily, unlike with the communities revolving around Microsoft and Apple, no one group can dictate what the rest of us do.  I am free to run with my idea, while others who are hesitant are free to stick with what they know and wait for the experimentalists to test the grounds before them.  This seems like the absolute best way to do things.

You seem to recognize that the base system is about as simple as it can get, or at least that this is the goal.  After that, you can just treat a single distribution as a client and you hardly get any additional complexity over a normal system (baring the init stuff, which still needs quite a bit of work).  After that, you just pull individual packages from other distributions as they are needed.  It really isn't that bad, at least once someone has reached a given minimum skillset.  It is weird to think of myself as some almighty Linux user who stands above so many others - I am sure there are quite a few others out there who are able to manage the system as I do.  I am doubtful it will ever get anywhere near simple enough to give distributions like Ubuntu or Linux Mint a run for their money, but within the circle of those who use, say, Gentoo, this is simple enough.

I do it the easy way: I don't do it at all.  I don't use anything which actually depends on dbus (I `chmod -x` the dbus executables so they don't pollute my process list), and I use mdev rather than udev.  PAM has yet to be an issue.

However, for those who who do want dbus and/or udev, the trick is to just use one from one distro at a time.  Many components which make up the modern Linux-based operating system are largely backwards-compatible.  All I have to do is make sure, say, udev, is sufficiently new to do whatever I want it to do.  The beautiful part about Bedrock Linux is that I can switch which distro's udev I am using on the fly.  If I want a cutting-edge one for whatever reason I can use that, or if I find it is broken for whatever reason I can fall back to the udev provided by another distro.

Yes, right now, this is an issue.  For others, anyways, as for me I hardly do anything on boot.  Before Bedrock Linux, every time I installed a package from my package manager which added some junk to what was launched at boot, I would manually go through and disable it each time.  Cups doesn't need to start at boot; I start it before I print something, print, then stop it immediately afterwards.  I like my process list to be nice and clean, and for everything to be as simple as possible while still providing me the functionality I need.

However, I would like Bedrock Linux to be used by those who actually do start things at boot which are installed by their package managers.  For the time being, for this to work with Bedrock Linux requires quite the user to manually configure parts of Bedrock Linux to start the various daemons from the client, manually specifying their order, manually parallelizing, etc.  This is certainly something I intend to improve eventually.  It could very well end up being an ongoing battle against ever changing client init systems at the rate things are going.

The brsh/brroot stuff is completely optional.  What it does is provide resiliency - if you would like to use a shell from a client that explodes for some reason, this is a none-issue in Bedrock Linux, as it will automatically fall back to the busybox /bin/sh.  As I said earlier, Bedrock Linux significantly eases maintenance for me.  I feel the complexity added by brsh/brroot is worth the it.  If you don't, but you also want to use Bedrock Linux - that's fine.  You don't have to.  Compile up your shell of choice and throw it into the core system, install it in all of your clients, and enjoy.

I do not think I missed that fact - do not necessarily always interpret my responses to anything you say as disagreement.  I just felt I had something to add given the exact item I was quoting.

For what it is worth, I initially intended this to just be a specific configuration for schroot, or perhaps a suite of tools built around schroot.  I later concluded that there would be strong benefit for placing this on top of as small of a base as possible, as described here: http://bedrocklinux.org/faq.html#Why%20 … tribution?

I am not sure if that FAQ item was up when this thread first started - it might be new to you.

After creating my own base for it, I realized that I created my own Linux distribution.  I did not set out initially set out with this goal in mind.  However, I did always intend on sharing what I had created should it prove useful to me - it just so happened that what I shared ended up being a distro.

And thus Bedrock Linux draws its strength from many sources

No need for concern there.  I put a very high priority on making sure my code simple, ease to read, and well documented.  For example, the 144-line "brc" utility has 34 lines including comments (and nine blank lines), and is described on multiple pages of the website.

You said you won't explain why you dislike Reddit, so I won't push you, but I'm really at a loss for how such a thing could generate such feelings.  I'm still relatively new to Reddit, so I could easily be missing something, but as far as I can tell Reddit acts more or less like a mailing list, with two main differences:  (1) it operates by "pull" rather than "push".  I get the latest posts to reddit.com/r/bedrocklinux when I go to check that website, not when it sends the messages to my mailbox.  I like that fact.  (2) It has a nice system to self-police.  This is a non-issue when traffic is low, but should things ever explode it could come in handy.  I recognize there are many issues with many of the subreddits.  I am free to avoid them - I can just go to the one for Bedrock Linux.  Just like with mailing lists, I do not have to subscribe to those I am not interested in.

I am perfectly fine with people expressing their dislike of Bedrock Linux (or really anything else).  I do agree that pretending there is value where there is not can have negative consequences.  I'd just ask that the reasoning behind it is explained, such that the actual value of the opinion can be taken into consideration.  If the opinion is baseless, or formed from insufficient knowledge, it is not assisting anyone any more than consciously overlooking problems.

You've given me one good thing to think about (and, should I have misunderstood your prior points, may yet provide me more in further responses).  However, I doubt it alone is sufficient to have me give this up; no worries there.  I'll almost certainly stick with this.  If nothing else, I want it for my own uses.  If other people like it, then I am happy to share.

Agreed.  Genetic variation is strength, contrasted strongly against monoculture.

Alas, should that day come, Bedrock Linux will go down with every else, as it has no legs of its own on which to stand.  However, if a only handful go haywire, those on Bedrock Linux should be relatively resilient to the consequences.

I appreciate your thought-out response.  So long as I understand why you feel the way you do about what I've done, even if I disagree with you, I am content with the situation.

EDIT: removed a typo

Hello, I am the lead dev behind Bedrock Linux.  I typically prefer not to jump in and discuss Bedrock Linux everywhere it comes up on the internet, as this would become an absurdly large task, and honestly I have no intention of convincing everyone of anything.  So long as there are some people out there who feel they benefit from this, I am content.  However, this thread has jumped up pretty high on the Google's rankings for "Bedrock Linux" and felt perhaps it would be worthwhile to make an exception.

Before I go into attempting to respond to everything said, I should state a few things:

(1) This is still well in alpha, and I don't propose anyone use this for serious production work at this point in time.  If you would like to help test and develop it, or to have a toy to play with, it may be fitting, but at the moment I do not necessarily recommend any more than that.

(2) Once the project matures to the point where I feel comfortable removing the above disclaimer, I still doubt I will consider it right for everyone everywhere.  There are many use-cases in which a traditional Linux distribution would be a better choice.  For example, the website for Bedrock Linux will be hosted on Debian for the foreseeable future.

(3) I am open to the idea that there could very well be faults with Bedrock Linux, both with regards to the specific implementation as well as the general concept.  However, I have yet to find anyone give a genuinely solid argument against Bedrock Linux's core idea other than "I, personally, have no use for it."  Most people who dislike Bedrock Linux for whatever reason seem to be operating under a false understanding of it - which could very well be my fault for explaining it poorly.

You seem to be operating under the assumption that the problem with Bedrock Linux is obvious, and haven't quite stated it exactly.  Would you mind explicitly doing so to ease further discussion?

Understandable, given the above statement.  However, your hope will come for naught, as this already has gotten a fair amount of attention and I fully expect it to continue doing so.  For what it is worth, the vast majority of the attention has been positive.

Unless you are using an extremely broad definition of "virtually," that is not true at all.  If you truly believe this, I strongly suspect either you did not read a sufficient amount of the documentation I have provided with regards to Bedrock Linux, or I did a poor job in writing the documentation.  For example, neither Gentoo nor Arch could provide an unchanging, reliable system for years as could be found with something such as Debian or a RHEL clone.

It is extremely tedious to do this with more than just a handful of packages.  Even more than that, in fact - eventually ones into problems with core libraries - such as glibc - having to be upgraded, which is not a fun project.

Primarily the latter, if I understand what you are trying to say.  However, I should note that two executables statically compiled at the moment with Bedrock Linux at the moment.

I never quite thought of it that way, but I that does make some sense.

I was trying to create something that fits my needs.  I've distro-hopped quite heavily and have yet to find a distro that does everything I want it to, so I made one that does.  It just so happens that what I created is extremely flexible and could very well end up fitting the needs of many, many others.

I do not follow quite why this is.

I will admit it is unusual, but I do not quite follow how it extends into the ridiculous.  Would you mind elaborating?

Well, this is true in the same way that one could install portagen another Linux distribution and claim it does everything Gentoo does.  The core concept could be, to a limited extent, used with other distributions.  However, to get the most out of it, one really needs a minimal base.  I was unable to find a minimal Linux distribution which was also well suited to be the base for a multi-user full-blown desktop, and so I made my own.

As I said earlier up, I do not feel this is anywhere near ready for a production system.  Once it does mature, however, I do not see why this would be a bad thing, so long as a capable system administrator was in charge.

Yup

Like Awebb and el_koraco, you seem to be operating under the assumption that the problem with this is obvious, however it is lost on me.  Could you explain exactly what is wrong with it?

If you have any questions or run into trouble when playing around with it, do not hesitate to jump into the IRC room or post on the subreddit.

I suspect you are missing two things: (1) for some people (such as myself), such items are not a rare case at all, and (2) while many are more than capable to build packages from source, we found it unnecessarily tedious.  I was finding parts of my Debian system were badly out of date on a regular basis.  I was wasting far, far to many hours compiling my own versions of things to fix them.  CUPS is to old for my new printer, the xorg-intel drivers are to old for my new laptop, the xserver itself is to old for some program I want, etc.  Trying a newer distro just resulted in other problems, as libraries ended up being to new and I had to play with LD_PRELOAD constantly to bypass the normal overly-new libraries, packages broke when I was far to busy with non-computer related items to allot the necessary time to fix them.  Maintenance over-all was far higher than I was happy to put up with.  I ended up feeling like I was more or less using Gentoo without the help of portage irrelevant of which distro I chose.

I eventually came up with a way to get everything I wanted from a distro: a very stable, reliable system, while at the same time fast, clean and easy access to newer arbitrary packages.  While I admit the initial setup is quite some work, for me, maintaining a Bedrock Linux system is absurdly easy.  Stable system to old to provide a package without compiling and dependency hell?  I just grab it from a newer distro.  Package in a newer distro breaks?  I just grab it from another distro.  Packages begin feeling disposable - I really don't care if they break.  I don't really have to debug anything anymore if something goes wrong on one of my computers - not that I couldn't if I wanted it, it is just that I don't have to.

Mind you I recognize that there are many out there who are quite content with a given Linux distribution, and for them I fully recommend they stick with it.  I just didn't find myself in such a category, and it seemed likely that there were others who felt similarly.

I hope I have clarified above that there is a far greater need for Bedrock Linux - at least for me - than just two packages a year.  I really do not find the system anywhere near "bable-ish" in complexity.  Fundamentally it is reasonably simple, just quite unusual.

That is all it really is at the moment, if I understand what you mean by that.  However, once it matures, I could see it being useful in servers and other areas as well.

I agree, all of those items could be solved without Bedrock Linux.  I used to run into such situations quite often before I came up with Bedrock Linux, and I was certainly knowledgeable enough to resolve them.  This issue is, resolving them without Bedrock Linux usually resulted in a bit of a time dump, where as with Bedrock Linux it became trivially easy.

This is an interesting point I've not yet heard.  In some sense it is certainly true - Bedrock Linux allows people to resolve issues without necessitating they know how to resolve the issues without Bedrock Linux.  I never really thought of this as a problem.  If an easy solution is available, why not use it?  My question, I guess, is where to draw the line.  I doubt you expect people to remove the cap off their CPUs and use a high-powered microscope to hunt down transistor-level issues, fix the design and have their processors re-fabricated.  Moreover, I doubt you'll find it acceptable that someone needs assistance debugging why a USB-peripheral which is not plugged in to a computer is not working.  Where between those two extremes is the acceptable medium?

I expected you had an issue with the core idea behind Bedrock Linux.  Since it seems based on this that you do not, I'm further confused about your issue with Bedrock Linux.  Is it that I am making it overly accessible?  Do you feel that people should pass some sort trial before being allowed to use this "nice thing"?

All the distribution does is automate a solution to problems I come across on a weekly if not daily basis.  I still do not quite follow why you feel it is necessary for people to use a more difficult solution when an easier one exists.

From this I gather you are attempting to say there is a easier solution and I am going about this an unnecessarily hard way, yet earlier I got the impression you felt Bedrock Linux was an overly-easy solution.  I suspect I do not understand exactly what you are trying to say.

I suspect you're trying to say something negative about Bedrock Linux with this, but I am at a loss as to what that point is.

For what it is worth, I've been using early versions of this for years.  I spent about nine months using what ended up being the first public alpha before releasing it without any serious issues.

If either of them - or anyone else - only finds two packages from which they feel they benefit from Bedrock Linux, then I fully agree they should go with something else.  The install time alone for Bedrock Linux would likely take more time than resolving two package issues.

Bedrock Linux's greatest strength, for me, is the incredible ease with which I can maintain the system.  If, say, I don't have time to follow the /usr move instructions in Arch, I can just dump all of Arch and use Debian Sid.  Maintenance is a breeze.  In fact, I thought you were trying to say earlier that it was to easy - perhaps I lost the intent of your message.

Due to the fundamental principle on which Bedrock Linux functions, it can never be as secure as a locked-down traditional distro.  If security is a high priority, I strongly recommend using something other than Bedrock Linux.

Yes, I admit, the knowledge level necessary to use Bedrock Linux is a bit high - you have to know the client distributions to some degree.  While I can get many aspects of other distributions through Bedrock Linux, I cannot get user-friendliness.  That is fine for me, as I am already quite comfortable with a variety of distributions - I became so when searching for one that actually has the features I want, before concluding I had to make my own.  This is not for everyone.

That did not come across based on your earlier statements.  Perhaps I misunderstood you - I am glad you feel that way.

It makes a fine desktop for me, but I already know both it and the client distros I use inside and out, and I have needs which no other single distribution I have found can alleviate.  While I recognize it will not be so for everyone, I fully expect there are others out there who could similarly benefit from Bedrock Linux as I do.

That just ended up being a side-effect; it wasn't really my initial intention.  I'm happy it ended up that way, though.

I don't follow why this would be if taken literally, nor do I follow how one should interpret it if it was meant to be sarcastic.  I am using reddit as a mix of what is traditionally being done through forums and mailing lists.  It seems to work fine for this thus far.

I feel as though the issues brought up were either due to false assumptions or misunderstandings, or that I was unable to properly guess what was meant and that I do not understand the issues brought up against Bedrock Linux.  If it was the former, I hope I have clarified the previously misunderstood issues.  If it was the latter, please do explain the problems again - perhaps re-wording them - so that I can resolve them.  If one of you are able to express to me clearly a serous problem with the core design of Bedrock Linux such that it cannot be fixed, I will be happy to abandon the project and take away from it what I have learned.  However, I am in doubt this will be the case.