Crime writer. Many of those crimes involve computers.
The FreeBSD Foundation has posted my BSDCam trip report. I also want to sincerely thank the Foundation for helping defray my expenses.
Robert Watson runs an excellent event. And what I learned at BSDCam will appear through the new Absolute FreeBSD.
Speaking of which, I need to go make some words. Bye.
I’ve talked to a whole bunch of folks who say things like “I’m a junior programmer. I’m looking for a way to help. I have no specific expertise, but I’m willing to learn.” Today, I present such junior programmers with an opportunity. An opportunity for you to learn skills that will be incredibly valuable to your career, and will simultaneously expand your career opportunities.
I’m at the FreeBSD BSDCam devsummit, gathering dirt information for the new Absolute FreeBSD. One of the topics is testing.
For decades, FreeBSD has relied on its users for testing. They expect users to install pre-release versions of the OS and exercise them to identify regressions. That’s necessary, but it’s nowhere near enough.
The FreeBSD Testing Project is building an automated test suite for the entire operating system. They have a whole mess of work to do. There’s only four people on the team, so each additional person that contributes can have a serious impact. They have tutorials on how to write tests, and sample tests.
There’s a whole bunch of tests left to be written. You have an almost open field. They need tests for everything from ls(1) to bhyve. (Yes, ls(1) broke at one point in the last few years.) Everything needs testing. Learning to write, submit, and commit small tests is valuable experience for developing the big tests.
And “I perform automated testing of my own code before the test team gets to see it” is a valuable item on your resume.
What’s more, learning to write tests for a system means learning the system. Developing tests will transform you into a FreeBSD expert. Once you’ve demonstrated your competence, worth, and ability to work within the project, other FreeBSD teams will solicit your help and advice. The Project will suck you in.
Testing is perhaps the most valuable contribution anyone can make to an open source project. And this door into the FreeBSD Project is standing wide, wide open.
For those who are watching me write the third edition of “Absolute FreeBSD,” here’s where things stand.
It’s at 77,400 words. 7 of 24 chapters exist [edit: oops, typo, not “chapter sexist,” spellcheckers are useless] as first drafts. Three more chapters are partially finished.
The current outline looks like this. Chapter titles subject to change. For reference, the ideal chapter length of a No Starch book is about 5000 words.
0 – Introduction (7500 words, d1 done)
1 – Getting More Help
2 – Pre-Install Considerations (4900 words)
3 – Installing (write last, because screenshots are evil)
4 – Booting (8300 words, SOL hates me)
5 – Backups
6 – Kernel
7 – Networking (8800 words, d1 done)
8 – Configuring Ethernet (7700 words, d1 done)
9 – Security
10 – Partitioning & GEOM (10700 words, d1 done)
11 – UFS (8000 words)
12 – ZFS
13 – Other Filesystems
14 – More Security
15 – /etc
16 – Packages (8600 words, d1 done)
17 – Ports (6800 words, d1 done)
18 – Software Management
19 – Upgrading
20 – Small System Services
21 – Performance
22 – Jails (6100 words, d1 done)
23 – Misc Crap
24 – Panics & Bugs
Afterword
Why write in this order?
When writing a large book, I always write the hard parts first. This means the book gets easier as it goes on. I don’t waste time or energy dreading That Topic.
It also means that I can write the really easy stuff when I literally cannot write the difficult stuff. I went to London, Ontario last weekend and was able to spend a few hours writing. I didn’t have my test lab, but I could sit on the front porch and write the Introduction.
Some of these chapters are difficult because of the research and testing. IPMI Serial over LAN is really ticking me off right now. I’m sure there’s a minor setting that I’m missing, and that once I find it everything will fall into place. Finding that setting will tell me what I don’t know, and let me explain and provide context. People buy my books for context. So, the product here is my frustration and suffering.
Some of these chapters are difficult because I’ve done too much research. I get a chapter on ZFS? Cool. But Allan Jude and I wrote two entire books on ZFS. Distilling that down to one-sixteenth the length, while providing context, and not shamelessly shilling those two books? Ugh.
Why these topics?
Because that’s what FreeBSD sysadmins must know.
Some of you will ask “where is bhyve?” Fair question. The bhyve developers are actively rototilling bhyve configuration. If I write a bhyve chapter now, it will be obsolete before the book hits print. That’s bad Including a bhyve chapter depends entirely on the bhyve devs.
If the bhyve devs settle on a configuration before the end of October, I’ll squeeze a chapter in after the jails chapter.
If they don’t, then you’ll have to wait for FreeBSD Mastery: Bhyve.
I’d like to have the third edition of Absolute FreeBSD in print by May 2018, for BSDCan. After discussions with No Starch Press, to make that happen I must complete the first draft before November 2017.
That’s four months.
To complicate things further, in those four months I have four weeks of traveling: BSDCam, EuroBSDCon, and a writing business workshop.
So, three months.
My best guess is that I have to write about 15,000 words a week to make that happen. Call it 3,000 words a working day. At my nonfiction cruising speed of 500 words an hour, that’s roughly 30 hours a week, or six hours a day. Provided nothing goes wrong.
I can only maintain that nonfiction speed if I write fiction as well. Without it, my nonfiction speed drops to about 200 words/hour. So I’ll keep doing 90 minutes a day on “Terrapin Sky Tango,” the sequel to Butterfly Waltz Stomp.
That’s 7.5 hours a day.
Plus I have a business to run. I have to assemble the “Core Concepts of ZFS” tutorial for EuroBSDCon. (If you’re in Europe and want to see me, this is your chance until some time in the 2020s.) I need to spend 2-3 nights a week on my martial arts practice.
The upshot is, you won’t see me around much until November.
The good news is, I’ve done the hardest parts of AF3e first. And I’ll squeeze in some time on weekends and evenings to get ahead.
But for now: book!
BSDCan ends each year with a charity auction. This year is no exception.
I’ll be bringing some unique items for the auction. First, a few things made by my wife. We’ll have a Relayd & Httpd Mastery keychain.
A FreeBSD church key. Yes, she made all the beads. They’re real glass. This is probably decorative use only.
A Puffy wine stopper. Again, all the beads are hand-made. Puffy is real glass.
SemiBUG is donating a Detroit sampler pack. By SemiBUG, of course, I mean me. Because my missus has all the hands-on talent in the family, and I thought I should bring something too. In any case, come by and visit our BUG some time.
Finally, I will have a small selection of 100% authentic, hand-crafted, truly bootleg Groff the BSD Goat merchandise. Get a piece while you can, I expect it to go fast.
See you all next week!
I finally have my grubby mitts on a print copy of Relayd and Httpd Mastery.
The cover is nice, but the inside has a couple problems. Specifically the TLS chapter.
I wrote this book on -current. Shortly after I finished writing about TLS, though, the -b flag for automatic backup was removed.
Someone pointed this out to me before I went to print. In my daftness, though, I misunderstood. (“Yes, yes, I know about the -b flag, what’s the problem?”)
There’s also a couple of other minor things–they changed acme-client -N to -D, and I missed one in an example. A couple sentences no verb.
These are embarrassing, but not critical if you know about them. The total error is about a third of a page, which isn’t bad for a 235-page book.
I’ve corrected the book. All ebook vendors have updated versions as of yesterday morning. Print versions ordered now should also be correct.
That leaves me responsible for three problems, though.
First, there’s no way I can have corrected print versions before BSDCan. It’s just not physically possible. So the copies I’ll have at BSDCan will have the error.
Second, 18 people ordered the print copy before I caught the error. Those 18 people are clearly both hard-core OpenBSD fans, and my hard-core fans. They are the people it’s most important I keep happy. It’s fully understandable that you’d want a correct book. If you desire a corrected print version, contact me. I’ll send you my mailing address. Mail me your incorrect book, and I’ll send you a correct one.
Yes, I’ll lose money on that exchange. Yes, I’m hoping most of you will decide to keep your “Special January 2017 OpenBSD-current Collectors Edition of Relayd and Httpd Mastery.” Yes, I’m hoping possession of that print book will turn into something that gives you hard-core bragging rights.
But you 18 folks? The ones who rushed out to get the book immediately upon release? You’re the ones I must keep happy.
So: the exchange is yours upon request.
If you’re one of those folks, do NOT return the book to Amazon! The Big A will put the incorrect copy on the shelf and ship it to someone else.
Third? Sponsors. Sponsors will get a copy of the good version. This means I must ship them after BSDCan.
Did this happen because I’m self-published? Nope. I submit the first sentence of Absolute OpenBSD as a counter-example.
Most of the time I don’t worry much about errata. I fix urgent daftness and move on. This one is particularly egregious and notably embarrassing, though–especially as it was pointed out to me before release.
The top of page 86 now reads:
find it easiest to create the backup file when create or renew a certificate. Back up the cert.pem, chain.pem, and fullchain.pem files. As I name each certificate file after the site, I can put all the backups in a location like /etc/ssl/acme/backups. Your update shell script needs something like this after a renewal.
DATE=$(date +%Y-%m-%d-%H:%M:%S)
cp /etc/ssl/acme/www3/www3.crt /etc/ssl/acme/backup/www3.crt.$DATE
cp /etc/ssl/acme/www3/www3.chain.pem /etc/ssl/acme/backup/www3.chain.pem.$DATE
cp /etc/ssl/acme/www3/www3.fullchain.pem /etc/ssl/acme/backup/www3.fullchain.pem.$DATE
Yes, anyone who bought the book could script better in their sleep. But the screwup is my fault and my responsibility.
I’m pleased to say that Relayd and Httpd Mastery is now available in print and ebook.
Sponsors should be able to log into their accounts and download the updated book.
I don’t have a print copy of the book in my hands yet, but they’re on the way. Unless something goes wrong, I’ll a) have copies mailed to all the print sponsors before I leave for BSDCan, and b) have a few copies for BSDCan.
Now to put in some quantity time for the third edition of Absolute FreeBSD before BSDCan…
Next weekend, April 28-30 2017, I’ll be at Penguicon. Two weekends after that (12-14 May), I’ll be at Kansas LinuxFest. But we’re on Penguicon right now.
Here’s my events and the description for each. Each is 1 hour unless specified otherwise. And I’m asking your help for some of these events. (Updated to add the LN2 events, which I’m not running but a guy has to eat sometime.)
Friday:
8PM: LN2 Ice Cream
9PM: The OpenBSD Web Stack – OpenBSD is best known for security and networking. But they also have a highly secure web server and load balancer. This talk will take you through the OpenBSD web stack, presenting its strengths and disadvantages. We’ll cover the httpd web server, free globally valid SSL certificates through ACME, the Common Address Redundancy Protocol for two-server clusters, and the relayd load balancer. Many of the security issues common on web servers are simply not an issue on OpenBSD. Come find out why!
Saturday:
9 AM: Writers and Traditional Publishing – So you want to sell a book to a publisher. How do you do that? What should you expect? How do you optimize your chances of getting not just a deal, but the deal you want? What gets some people into traditional publishing, and keeps others out? Come hear authors discuss the good and bad of the publishing biz!
10-11:45AM: Author Meet & Read, Vol. 1 – A big room with Clif Flynt, Mary Lynne Gibbs, Jen Haeger, Christian Klaver, James Frederick Leach, David Erik Nelson, John Scalzi, Clarence Young, and myself, all showing off our books, talking to our readers, and signing books. I will have my books still in print for sale. I’m expecting that the others will all have long lines and I’ll be there alone, so this is your chance to heckle me in person.
10:54-11:03AM: reading from git commit murder – Readings are tightly scheduled, so I expect this to begin and end sharply on time.
1PM: self publishing in 2017 – Self-publishing is an increasingly important channel for authors to reach their readers. It also changes constantly, with new tools and distributors opening daily and existing platforms changing. This panel brings together veteran self-publishers to share their experiences, discuss the changes of the last year, and give new authors an edge in the business.
2PM: 90 second reads – Join a handful of Penguicon authors as they read 90-second passages from their novels. The selections will be thematically linked based on keywords, such as sorrow, fury, funny, love, etc. Timing is crucial! After, there will be a Q&A with the authors.
3PM: LN2 ice cream
5PM: Writing High-Performance Nonfiction – Writing nonfiction is not merely reciting facts. It’s a specialized form of storytelling, very different from your college essays and book reports. Whether you’re writing memoirs or computer texts, using storytelling techniques transforms your work for the better. This talk takes you through making your nonfiction not only readable, but memorable.
7PM: BSD Operating Systems in 2017 – I’ll be discussing the current options in BSD-based operating systems, the big news from recent projects, new developments, and where we’re going from here.
8-10PM: LN2 ice cream
Sunday:
10AM: breakfast – LN2 ice cream
11AM: Senior Sysadmin Panel – Storage – The years know things that the days and weeks never know. We’ve gathered half a dozen people who’ve been sysadmins for over 20 years to talk about the one of the most dreaded and annoying topics in computing: storage.
12PM: Self-Promotion for Creatives – Independent creators are their own PR departments. We have to not only make all the things, we have to spread the word about all the things. Here we have a bunch of artists and writer types who successfully spread their work across the world. What works? What doesn’t? How can you be shamelessly self-promoting without being a jerk? Come find out!
Where could I use help?
In the 90 second reads panel, I get a few 90 second periods to read a selection from my fiction. Each read should have a theme. Our group has four themes: Betrayal, Heartbreaking, Scary, Funny.
For those of you who have read my fiction: I could use suggestions for parts of my books that you thought fit these themes. I have a few thoughts, but what I think fits a theme is probably not what struck you lot as fitting that theme.
So: if you’ve read my fiction, what of mine would you suggest for a brief reading in any or all of those themes?
Reyk got his tech edits on the “Httpd and Relayd manuscript back to me. I considered sending it out for a wider tech review, but it’s been through a few hands already. And nobody knows this software better than Reyk.
So I’ve sent it on to copyedit.
If you’re one of the folks who have a copy of the manuscript: I can use any feedback I get in the next couple of weeks.
The book won’t be ready for Penguicon, sadly. I should have it in print for BSDCan, however.
I will have stickers of the front cover for BSDCan, though. And stickers of the git commit murder cover.
I’m pleased to say that with a bid of $800, Bert JW Regeer has won the Absolute FreeBSD 3rd Edition Tuckerization Auction.
Bert, send me your donation receipt.
After the Httpd and Relayd Mastery Tuckerization auction, I’m slightly surprised that this auction didn’t go higher. AF3e is a bigger tome than the relayd book. I’m forced to conclude that the FreeBSD folks just have less team spirit than the OpenBSD folks. (Not zero, mind you, but 8/13th as much.) Or the OpenBSD folks are accustomed to paying for things.
Or, just maybe, FreeBSD fans know better than to spend their hard-earned money on silly auctions.
Nah, we’ll go with team spirit.
Thanks for playing, everyone. While these have been fun, I don’t intend to do any more Tuckerizations for a while. The gag gets old quick.