Bjørn Stabell 白熊

In his article Jacky explained how to easily drill holes through to ports on machines behind a firewall. What I normally want is to have ssh access to machines behind firewalls, allowing me to do scp, and easily ssh in without a stupid stop-over on the firewall machine.

I came across this solution that does exactly that. After the super-simple set-up I’m able to do:

All that’s required is to tweak your .ssh/config. Mine looks like this on my laptop:

Warning: This is a rant.

With wifi came the promise of being online (almost) anywhere, but due to incompetent or misdirected implementation and management, it’s pretty much a patchwork of extremely unreliable networks. My experience is that there’s a 30-40% chance of actually being able to get online at an access point.

(more…)

For a while now I’ve been aware of this situation, but thinking it would pass I didn’t do anything. Well, it hasn’t, so I thought I’d start sharing what I see.

Beijing Telecom has broken the Internet

At various places, one or more of the DNS servers you get from, e.g., wifi hotspots, will be configured to always resolve any domain name lookup. That means typos like www.doesnotexist11.com will resolve.

What do they resolve to? They resolve to IP addresses for websites of advertisers. It’s most likely a pay-per-referrer business model.

(more…)

Although it’s relaxing and nice to visit relatives in Norway, I’m starting to look forward to EuroPython 2006 next week in Geneva, Switzerland.

Exoweb is going with three people: Ken Wong, Gzregorz Jakacki, and me.

Are you going as well? Drop us a line. Would be fun to meet up and chat, especially if it’s about Python, Django, Dojo, or Postgres.

For many years I’ve been in companies that, sometimes thanks to my stubbornness, were using Debian GNU/Linux as the general office server. As the number of services and the integration requirements (we don’t want 10 disjoint user databases) of these services have grown, this has become quite painful; Debian provides little help with such integration. Here is a list of sample requirements:

• Firewall with NAT
• DHCP server connected to DNS server
• Windows file sharing (Samba)
• Apple file sharing (Netatalk)
• Linux file sharing
• SMTP server with authentication, spam and virus checking
• IMAP server
• VPN server
• Intranet services such as Wiki and Issue tracking
• Shared accounts across Linux machines
• Shared calendaring and mail folders a la Exchange (if feasible)

All of these services should be operating from one common user database, pre-configured to work together, and continue to work flawlessly as Debian packages are updated. It should fit the requirements of small and medium-sized enterprises/ businesses (SME/ SMB) as much as possible.

A couple of years ago, I became aware of the e-smith server, aka Mitel SME Server (information also on contribs.org) that does a great job at the above, but it is a derived Red Hat 7.3 distribution, old and relatively useless by now. Looking around I could find one other request for a SME server equivalent for Debian, but Alex Polite had not gotten any replies when I asked him. If anyone know of anyone doing such undertaking, please let me know.

Assuming there is no project underway to make it such a LAN server, I would like to start gathering requirements for such a server, to be implemented as a Custom Debian Distribution (a meta package), allowing anyone with a stock Debian distribution to be able to easily configure all the above services, while still enjoying the ease-of-upgrading and wide base of packages available for Debian. (See an earlier post about not reinventing the wheel for more benefits of this approach.)

If you’ve got requirements, general thoughts, information, or want to participate, please reply.

[ See discussion about this on debian-custom. ]

I love free and open source software, but there’s one characteristic of the FOSS community that is driving me mad: there are just too many me-too projects! Instead of cooperating and solving new problems, most seem to set out to solve the same problem again and again.

Case in point: It’s just too easy to make a Linux distribution these days–witness the top 100 distros at distrowatch. It seems every kid on the block is creating a distribution, with nothing original except it’s got their own my-name-ix. This has many problems:

• Software vendors can’t support 100 Linux distributions, it’s just too much work
• The package maintainance, system development, and support communities are fragmented, failing to reach critical mass on certain topis and problems
• New users are confused with too much choice, too little direction, and may become very disillusioned if they happened to chose a distribution that has a dead end (very likely since many distributions are just one-person operations)

I’m not saying innovation isn’t good, but there are better ways to innovate so that FOSS is strengthened, not weakened. Let’s say you’re a Linux geek working in the insurance industry, and you’d like to improve Debian for people like yourself:

• You might chose to create a Custom Debian Distributions (CDD) for the insurance industry. This is not a Debian fork (derived distribution) a la Ubuntu, but a meta package which specifies a set of packages, customizes them, and configures them to work together in a certain way. In this case, the “debian-insurance” CDD would probably select a few packages that might be good for insurance people and configure the applications menu with only relevant menu items. You should, of course, also create a community that allows other insurance people to participate in the continued definition and maintenance of this new CDD. This is as easy as creating a Source Forge project.
• You might also chose to become the Debian maintainer for a few insurance-related packages that are not in Debian already, packages that aren’t duplicating functionality already available.
• If you find some problems with packages you want to use, let’s say they can’t easily be configured the way you want them to be, you should provide patches and extensions to the upstream Debian package maintainers so that everyone can enjoy this new functionality or bug fix.

This has some wonderful benefits:

• You’ve added value to the base Debian distribution, making FOSS and Debian stronger instead of weaker, as well as simplifying the choice of the end user.
• You’ve minimized my own maintenance burden, since you can automatically benefit from upgrades of other packages, extensions to Debian, etc.

To summarize my plea to all FOSS activitists:

Don’t re-invent the wheel.
Don’t let shyness or laziness stop you from connecting with and joining other communities out there that are also focused on solving the problem you want to solve.
Forking is the last and worst option; it’ll increase your maintenance burden and weaken FOSS overall. You are in essence creating a competitor.

This me-too problem applies not only to distributions; it’s easy to see the same effect in, e.g., web application frameworks and language modules as well.

Or as programmers like to say: Hello World! After spending so much time investigating the blog engines out there I thought a “vrooom” might be a better title for this, the start-up post.

Although my favorite technical environment would be Django and Python, and Socialist Software has made a decent implementation of the WordPress K2 theme using Django, I still opted for a real WordPress installation so I can get used the the full power of a mature blog software. At least this way, when the itch of not running on Django gets too noticable, I know how to scratch it

I plan to write on subjects that matter to me at any given time, syndicating the whole or parts of the site (using tags) to other places. This include:

• General business, management, agile software development, and human resources issues relating to Exoweb, the outsourced software research & development firm I founded and am still working in. This will be syndicated to the Exoweb planet site when it is ready.
• Technology issues, generally open source, and mostly relating to web application development using Ajax, Django, Python, or Ruby on Rails. A post or two on Mac or Linux might sneak in as well, as I’m quite fond of the Unix family of operating systems. These posts may be syndicated to other technology sites once I figure out which ones and how to do it.
• Impressions on the incredible place I live in–Beijing, China–or memories of the even stranger place I come from–Havøysund, Norway. It’d be nice if there were some RSS aggregators for these communities.
• Ideas on personal development and lifehacking. I’m blessed with good sense of logic, but cursed with bad long-term memory. Without systems I just can’t deal with all the stuff thrown at me.
• Personal experiences that I want to share with friends and family. Pictures will be posted to my beorn flickr account.

Don’t hesitate to get back to me with comments. I expect the more feedback I get, the more I’ll write. We humans are funny that way.

Thanks for reading! This blog is on the roll…!