Recommend: A simple GUI for OpenBSD install

[cravuhaw2C]

I wish to recommend that OpenBSD developers create a simple GUI to help novice users to install the OS.

If anyone had opted to use the "Expert Install" option of Debian, you would have realized how easy and intuitive the simple GUI was.

Special note: if OpenBSD developers plan to take up my suggestion, please simplify the process by which one partitions the disk drive. Debian's "Expert Install" option does not even mention C/H/S or LBA. When I was using Debian for the first time, I was able to partition my disk drive without problem.

[ibara]

No.

[jggimi]

Almost all of the OpenBSD users here are end-users. And we have learned over the years that the social culture of the project needs to be understood by new users:

• The goals of the Project are to meet the needs of its members, the developers themselves.
• We users come along for the ride. If we like what is offered, we are free to use it. If we don't like it, we are free to look elsewhere for systems, services and tools. However...there is some balance: if the Project's software does not meet users' needs, Project revenues (CD sales, donations) will decline.
• If users have suggestions for improvements, they will be welcomed and reviewed for acceptance if:
  • They are accompanied by users' own development efforts
  • For hardware support suggestions, they may be accepted if they are accompanied by donated hardware for R&D and if a developer has an interest. Both are required.
• Most users who become developers -- Project members -- have begun by providing patches or developing components they feel would enhance the OS. Most would tell you that their first attempts were rejected, even rejected many times.

The installation script that you find so offputting is liked by all of the Project members, and by most of us users. If you were to design a graphical installer, you are unlikely to have it accepted.

(And there are many platforms that have no graphics at all. I mentioned one to you in another of your many threads: an Alix. I run OpenBSD on several.)

[cravuhaw2C]

Originally Posted by jggimi

The goals of the Project are to meet the needs of its members, the developers themselves.

Thanks for introducing the social culture of OpenBSD to me.

Originally Posted by jggimi

We users come along for the ride. If we like what is offered, we are free to use it. If we don't like it, we are free to look elsewhere for systems, services and tools. However...there is some balance: if the Project's software does not meet users' needs, Project revenues (CD sales, donations) will decline.

From what I read and understood, based on the articles available via Google Search, OpenBSD aims to be the most secure *nix OS on the planet. Does OpenBSD aim to be the most popular *nix OS?

Life is getting harder as we speak. Unemployment rears its ugly head from time to time. Recession cycles are getting shorter: big recessions occur once every 5 to 6 years in our lifetimes compared to the 1930s to 1990s.

OpenBSD developers have lofty ideals but putting food on the table is the overarching primary concern. Maslow's hierarchy of needs applies to each and every one of us.

I read on the internet that last year OpenBSD threatened to stop development of its project because it was running out of funds. Subsequently it received about $150,000, only for this year, I think? What about next year and the years following it?

Debian, Ubuntu and Redhat do not have funding issues. OpenBSD should ask themselves why.

IMHO aspiring to be the most secure OS does not contradict with being the most popular *nix OS.

What do you think?

Originally Posted by jggimi

The installation script that you find so offputting is liked by all of the Project members, and by most of us users.

I apologize if I came across as finding the installation script offputting.

I was OK with it till the part where I have to partition my SSD for a multi-boot environment. I'm stumped by all the talk about cylinders, heads, sectors and sector geometry. Debian and Ubuntu took care of that for us who are technically or mathematically challenged. Why can't OpenBSD do the same for us?

I was a bit shocked by the advisory in the official FAQ in which first time users are discouraged from installing OpenBSD in a multi-boot environment. No such advisory exists for Gentoo, ArchLinux, Mandriva, Debian and Ubuntu. I now understand why. It's because users are required to have good math skills to be able to calculate C/H/S or LBA sector values.

If OpenBSD is to gain a wider acceptance among the public, it has to find ways to make the installation routine more user-friendly, or should I say, more novice user-friendly.

Note: by "wider acceptance among the public", I'm thinking along the lines of "having sufficient funds to keep the OpenBSD going for a few more years".

[blackhole]

Originally Posted by cravuhaw2C

If anyone had opted to use the "Expert Install" option of Debian, you would have realized how easy and intuitive the simple GUI was.

The "expert" install is actually the only useful installer. It's also dead easy for new users. The GUI installer is a pointless duplication of an already working installer with less configuration options. The installer is just there to get the OS installed, after that you don't want to see it. The OpenBSD installer works, so I don't see the problem, nor see why the comparison with other OS' installers is relevant.

You need to stop comparing and discussing other operating systems. If you like how those other OS or Linux distributions do things then you have the option to just use Linux. I don't see the problem.

Originally Posted by cravuhaw2C

mathematically challenged.

I'm severely mathematically challenged, yet somehow managed to create two partitions. On my first install of 5.4 I went for the auto partitioning, but that did not leave me with enough space on / so I reinstalled and had to sit in front of it for about half an hour read the manual and work out the basics of fdisk's interactive mode. Somehow I managed it. When I installed 5.5, I had to relearn it as I'd forgotten how to do it...

Originally Posted by cravuhaw2C

I was a bit shocked by the advisory in the official FAQ in which first time users are discouraged from installing OpenBSD in a multi-boot environment. No such advisory exists for Gentoo, ArchLinux, Mandriva, Debian and Ubuntu

Yet those distros' forums are not awash with new users who have accidentally hosed their systems and nuked their windows into oblivion....?

Originally Posted by cravuhaw2C

I now understand why. It's because users are required to have good math skills to be able to calculate C/H/S or LBA sector values.

No this is not the reason.

Originally Posted by cravuhaw2C

If OpenBSD is to gain a wider acceptance among the public, it has to find ways to make the installation routine more user-friendly, or should I say, more novice user-friendly.

So if the installation routine is dead easy even a monkey can do it... then the set up of the OS and installation of software requires some skills/work on your part - as is currently the case - what then? Why have an easy installer if you're not also going to have an easy OS?

[jggimi]

Originally Posted by cravuhaw2C

Does OpenBSD aim to be the most popular *nix OS?

No. Unlike every Linux you have ever used, or any commercial OS, the Project is very, very different. It has absolutely no interest in being the OS-for-everyone. None. If you don't like what they offer, you are welcome to go away. If you read through the rough-and-tumble archives of the Project's misc@ mailing list, you can see many examples of this latter suggestion.

The "market" for users of interest to the Project, such as it is, does not include non-technical, casual computer users.

And that is why the openbsd-newbies mailing list and this forum are both unofficial.

OpenBSD developers have lofty ideals but putting food on the table is the overarching primary concern.

Nope. This is a non-commercial endeavor. Project funding covers infrastructure, select travel, and a single salary.

I read on the internet that last year OpenBSD threatened to stop development of its project because it was running out of funds. Subsequently it received about $150,000, only for this year, I think? What about next year and the years following it?

You are correct. Infrastructure operating expenses had skyrocketed. But that does not mean the Project is considering broadening its focus to include non-technical users.

I was a bit shocked by the advisory in the official FAQ in which first time users are discouraged from installing OpenBSD in a multi-boot environment.

Even if you do the math right, the bootloader is dependent on BIOS services initially. I referred you to FAQ 14.8, and now do so again. You may be able to install OpenBSD anywhere on a drive. But your BIOS may prevent its being able to boot, depending on the BIOS and where on that drive its second stage bootloader is located. Can this be circumvented, if this happens? Only by relocating the OS.

If OpenBSD is to gain a wider acceptance among the public, it has to find ways to make the installation routine more user-friendly, or should I say, more novice user-friendly.

As mentioned, these users are welcome to become less novice and more self-supporting, or they are welcome to hire a commercial service to provide the end-user support the Project cannot provide and is not interested in providing. The Project's website lists commercial service providers for this reason.

[Oko]

Originally Posted by cravuhaw2C

OpenBSD developers have lofty ideals but putting food on the table is the overarching primary concern.

This is factually wrong. OpenBSD licensing model is copy right and copy left licenses GPL for example are avoided by all means.

Originally Posted by cravuhaw2C

Debian, Ubuntu and Redhat do not have funding issues. OpenBSD should ask themselves why.

Red Hat is fortune 500 company publicly traded on the NY stack change. Canonical (Ubuntu) as well. Debian is the largest hobby project on the world. Yet amount of innovation which came out of 100 something OpenBSD developers exceed these three companies combined together.

Originally Posted by cravuhaw2C

Note: by "wider acceptance among the public", I'm thinking along the lines of "having sufficient funds to keep the OpenBSD going for a few more years".

You do not have sufficient funds to support OpenBSD. People who have like Google have no problems installing OpenBSD. Recently the offer all their in house modification of OpenSSL back to LibreSSL and change the license to the same one as OpenBSD.
30 000 elevators in U.S. run OpenBSD as do majority of network switches and Internet infrastructure. That seem like a pretty good adoption rate in my book.

[marcolino]

Originally Posted by Oko

30 000 elevators in U.S. run OpenBSD...

That is one heck of a statement, and one I've never heard before. I'm intrigued - do you have more information about this, or perhaps a link to share?

(Sorry for the off-topic post.)

[Oko]

Originally Posted by marcolino

That is one heck of a statement, and one I've never heard before. I'm intrigued - do you have more information about this, or perhaps a link to share?

(Sorry for the off-topic post.)

There was a thread few years back on misc when somebody was nagging basically claiming that nobody uses OpenBSD because of unfriendly mailing list. One of the guys had it and just opened up and stated bluntly that he runs OpenBSD on 30000 elevators. I checked the name of the people made that post and sure enough the guy is the president of the company which does indeed run large numbers of elevators in U.S.

[spitfire_ak]

Originally Posted by jggimi

Almost all of the OpenBSD users here are end-users. And we have learned over the years that the social culture of the project needs to be understood by new users:

The goals of the Project are to meet the needs of its members, the developers themselves.

Then the OpenBSD system is not suitable for production environments; as an example, AIX listened to its customers, my employer, and developed suitable options to fit within IBM's AIX framework to meet those customers' needs; an outright "No" is hardly a great way to get others to use OpenBSD.

Also, if OpenBSD's goal is to meet ONLY its members' needs, why release it at all? I think in releasing something to the public entails collaboration with that public; if not collaboration, a patient response to why something is not done or will never be done.

I am new to OpenBSD and not exactly new to Unix/Linux; overall, the install procedure is not difficult and, yet, neither is it intuitive. Especially when trying to follow the automatic partitioning of slices... I have 16GB or RAM and the auto partitioner allocated 16GB of swap space! In 1995 swap should mirror RAM; in 2014 that might not be the case for almost all modern systems; further, the default list in disklabel slices is in some esoteric format not spelled out by disklabel(8) or fdisk(8). In the absence of options in 'p' (i.e., [b]ytes, [m]egabytes, [g]igabytes, etc.), what is the default list unit? No math I've found makes sense.

I've noticed that OpenBSD attracts a lot of "RTFM" responses; the problem with that is that man/info pages are written by those who know for those who know but may need a little refresher on flags, etc. They are not friendly to new users... I would never, and I am assuming none of you would do the same, EVER point my wife to disklabel(8) and tell her to RTFM, figure it out and good luck. She would stick with Mac OS X or Linux rather than have to slog through the developers' esoteric man/info pages; she has tried and, being dyslexic, found it extremely difficult.

Personally, though, I thought OpenBSD's install procedure to be great, and aside from a little more learning on my part, easy compared to other operating systems. Whole disk encryption is a breeze compared to most Linux distributions!

[jggimi]

Originally Posted by spitfire_ak

Then the OpenBSD system is not suitable for production environments; as an example, AIX listened to its customers, my employer, and developed suitable options to fit within IBM's AIX framework to meet those customers' needs; an outright "No" is hardly a great way to get others to use OpenBSD.

OpenBSD doesn't have customers. It has users. And users are "listened to" primarily in four ways:

1. bug reports
2. donations of money or equipment
3. CD purchases
4. development efforts via user-provided software or patches

Also, if OpenBSD's goal is to meet ONLY its members' needs, why release it at all? I think in releasing something to the public entails collaboration with that public; if not collaboration, a patient response to why something is not done or will never be done.

This is the official goal list:

http://www.openbsd.org/goals.html

I am new to OpenBSD and not exactly new to Unix/Linux; overall, the install procedure is not difficult and, yet, neither is it intuitive. Especially when trying to follow the automatic partitioning of slices... I have 16GB or RAM and the auto partitioner allocated 16GB of swap space! In 1995 swap should mirror RAM; in 2014 that might not be the case for almost all modern systems; ...

I disagree with both of these opinions but this may just be because of longer experience. I find the installer both easy and intuitive, and my swap space is at least larger than main memory size on non-embedded platforms, as I wish to be able to obtain and analyse dumps from kernel crashes. These are stored in swap space, uncompressed, for later storing in /var/crash after reboot.

... further, the default list in disklabel slices is in some esoteric format ...

the default is sectors, as explained here. Both programs allow the user to use bytes, kilobytes, megabytes, gigabytes, terabytes, and cylinders for partition start and partition length values.

I've noticed that OpenBSD attracts a lot of "RTFM" responses; the problem with that is that man/info pages are written by those who know for those who know but may need a little refresher on flags, etc.

The "info" pages are only applicable to applications obtained from the GNU project, and as they are not maintained by OpenBSD, anything confusing or misleading in them should really be reported upstream.

The OS's definitive documentation is its collection of man pages. However, they are not intended to be tutorials.

The only official "How To" documents are the FAQ and its subsidiary documents such as the PF Users Guide.

[ibara]

Originally Posted by spitfire_ak

Then the OpenBSD system is not suitable for production environments

I like you. You're funny.

[spitfire_ak]

See: Comment from Mr. Raadt ... OpenBSD is in very REAL trouble as of 2014. Hopefully, some company or person has stepped up to provide the much-needed funds.

I have a little money to donate, I paid for subscriptions to SUSE and am now moving on. I'm waiting to see how I feel about OpenBSD before investing more time & money.

As for the OP, he asked a legitimate question: why no GUI install? Heck, I would go further, why no curses-based install, at the least? Is there a security reason or just a preference by the developers?

[ibara]

Originally Posted by spitfire_ak

See: Comment from Mr. Raadt ... OpenBSD is in very REAL trouble as of 2014. Hopefully, some company or person has stepped up to provide the much-needed funds.

I like how you're presuming to tell me things. It's cute.

Originally Posted by spitfire_ak

As for the OP, he asked a legitimate question: why no GUI install? Heck, I would go further, why no curses-based install, at the least? Is there a security reason or just a preference by the developers?

OK. Let's say you have to install OpenBSD hundreds of times a year, on a wide variety of system architectures (yes, I do exactly this. If you need to ask "well why would you do that?" consider clicking on the link in my sig instead of wasting time pondering it): would you rather have a GUI that is likely unmaintainable across all the architectures (especially since there are many machines that, you know, don't have VGA. the world isn't all a PC these days... never was, really) or a curses based installer that is ugly over serial (and again, more complex than a simple shell). Or would you rather have a streamlined shell that does everything you need, doesn't get in your way, and can almost literally be installed by holding down the Enter key?

If you're looking for complexity and nonsense, OpenBSD isn't the project for you. Really. And that's OK. OpenBSD isn't the project for a lot of people. But now your extrapolations have gotten beyond the realm of silly.

If a shell installer is something that's so major you wouldn't want to install OpenBSD, then don't install OpenBSD.

Or you can stop complaining and do something about it. Worked well for me.

But really, the GUI installer is something that's been asked time and time again. It's not gonna change. It's tiring to see people argue for something they haven't even attempted to understand.

[hitest]

Originally Posted by cravuhaw2C

I wish to recommend that OpenBSD developers create a simple GUI to help novice users to install the OS.

Uh no. The OpenBSD installation routine is perfect as is. I would not want to see a GUI installer on my favourite BSD.

[jggimi]

In follow up, I was going to send a private message (PM) but you don't have them enabled for your account. So I'll post it here, since none of it is actually private.

As an example of commercial, end-user support services ....

One of the services firms listed on the Project's Commercial Support page is M:Tier Ltd. They are particularly active in support of the OpenBSD community at large, and offer a number of free services such as binary packaging for OpenBSD-stable users.

I mention them because their commercial services include ready-to-use Gnome workstation enviornments, something you expressed interest in. And regarding Gnome, their staff includes Antoine Jacoutot (ajacoutot@), who leads the OpenBSD Gnome porting team among his many other accomplishments for the Project.

I don't know if they would be interested in supporting individual customers rather than their usual corporate fleets -- but you could certainly ask, if interested.

http://www.mtier.org/solutions/os/openbsd/desktop/

[JWJones]

Perhaps you should consider hardening your Debian installations, rather than trying to dumb-down OpenBSD for the masses of Linux distro-hoppers. You will find that most BSD users don't give a damn about Linuxisms and concessions towards "user-friendliness."

[Oko]

Originally Posted by JWJones

You will find that most BSD users don't give a damn about Linuxisms and concessions towards "user-friendliness."

I agree on Linuxisms part but I vehemently disagree with your second claim. We really, really care about "user-friendliness". BSDs are some of the most user friendly systems in existence. The only problem is that BSDs are very peculiar about the choice of their friends and users.

[JWJones]

^ Haha, yes, good point!

[blackhole]

"user friendly" is a marketing term. And that's really the whole problem. From an individual's perspective it actually means: "how my other OS worked [and how I expect others to work]".

Most of the common proprietary GUIs are marketing devices - in that they strive to be distinct from the competition and provide something new/different just "because". Usability is generally sacrificed because the GUI has to make an impact, be "innovative" and grab the headlines - hence windows 8 "metro". These kinds of duff decisions are usually shareholder driven.

The command line is user friendly to a person who knows how to use it. A clunky GUI, especially with huge buttons designed for touch screen may not be user friendly to everyone. A lot of people found gnome 3 user friendly, a lot did not, etc.