What should I write?

I have always wanted to write (English) more; turns out that in order to write you need something called “time.” I have finally succeeded in carving out time to write (thank you, Red Hat) and now I have the luxury of wondering what to write. I have my own ideas, but perhaps you, gentle reader, have some thoughts on the matter?

28 thoughts on “What should I write?”

  1. Mistakes

    I always find it interesting to read about mistakes others have made. It is a luxury to be able to learn and avoid other’s mistakes :)

    Another suggestion is alternatives to the conventional. Think “Innovator’s Dilemma” where small insignificant things eventually diminish the big (and usually profitable) stuff that is already there.

  2. Re: that’s easy

    Heh. Adding an extra layer of indirection, I also remind him to remind you to write more LWN articles every so often :-) Does that mean me writing this comment counts as a layering violation?

  3. Try sci fi – maybe the ‘hard’ one (i.e. more on the ‘sci’ side, in your case being kernel programming). As far as I can tell, you should be really good at it.

  4. Hello Val,

    I’ve recently encountered a problem with fs free blocks map. I needed to get a said map and project it to block device 512-bytes physical blocks. The actual task was to port a windows implementation of virtual crypto-loop device driver and user space app. It has a specific container file format that stores service info and actual encrypted data formatted in ntfs (most likely) that gets mounted as a virtual device. This container also can have a “hidden part” that answers to certain password\public key and is being stored in the original container filesystem’s free blocks. When this part is mounted the driver gets a block map to represent as a contiguous space resulting in a separate slice of data. To create said map in the first place i needed to know which ntfs block are free. And i’m still can’t quite figure out if this is at all possible to do on Linux without relying on ntfs (or any other supported fs) specific details. Now i’m not asking you to write an article about it. But *maybe* some of this is an interesting material for you.)


  5. Write more about linux desktop experience.

    — Any thoughts on the subject of how pulseaudio developers are “fixing” “linux audio mess”? From my own anecdotal evidence open sound system feels more instantaneous and lighter on cpu while mixing audio and controlling volume per application.

    — Why would any hard drive manufacturer choose insanely aggressive load/unload cycle values by default?

    — Why doesn’t my external hard drive spin down after I click “unmount” in Ubuntu, continuing to make noise until I give up waiting for it to enter low power mode and unplug usb connector? Windows does it momentarily after choosing “eject”.

    — What filesystem do you personally use on usb flash drives?

  6. Um…

    What kind of writing do you want to do? If it’s technical writing, well, I could use a book on the general theory of modern file systems on my shelf :-) Or maybe you’d like to go in a different direction and write some fiction? So much to choose from.

  7. About filesystems

    Hello Val,
    I read your blog since a few months and I saw articles about filesystems. If this doesn’t seem boring to you, I’d be happy to read explanations about the principles of filesystems, the differences between some of them, things theoretic about them…


  8. filesystem internals

    More about filesystem internals, and the tradeoffs involved in designing a filesystem. Things like code complexity, API complexity, obvious space/time tradeoffs, practical things like limitations imposed by the OS, etc.

  9. What you’re great at of course

    How about a history of how you got into kernel hackerage? Its always interesting to see the bumps along someone else’s road as a way to swerve around them on your own journey.

  10. I would like to read reviews of fiction you read/movies you see.

    If you want a specific LWN kernel request (and you probably don’t): log structured file systems! I’ve read Rosenblum and Ousterhout, “The Design and Implementation of a Log-Structured File System,” and I suppose I could just read more of the literature, but isn’t it easier to sit on my hands and patiently wait for you to summarise it for me? Yes!

    Otherwise my general preference for non-fiction reading is revealing the hacker ethic everywhere. People obsessing about problems I hadn’t ever heard of. People and problems, my two favourite things.

  11. Re: Write more about linux desktop experience.

    — What filesystem do you personally use on usb flash drives?

    ext2. And I always run cksum on anything important, before and after, because data corruption is almost the norm for USB sticks. (I don’t use md5sum; it takes much longer to run and all I’m looking for is error detection.)

    I am shamelessly dodging all your other questions.

  12. Re: Mistakes

    > I always find it interesting to read about mistakes others have made.

    I’m afraid that “mistakes I have made” is rather too broad a topic to cover…

  13. Hm, you know, I was hoping someone would say, “Please write science fiction” when I posted this, and yet I would be embarrassed to actually try. I will ponder.


  14. Re: What you’re great at of course

    Unfortunately, I don’t think the history of how I got into kernel stuff would be terribly instructive… I got lucky enough to take Operating Systems the year that it was taught using Linux – real mainline Linux running on real hardware. Then I saw an ad in the Albuquerque Journal “Help Wanted” section for a Linux kernel developer for a local network hardware company.

  15. Or take the Nano approach

    Don’t think. Just write. Whatever comes out. Write as if you’re on fire! Write furiously! Write! Let it it all come out and when it stops, knock yourself unconscious so your internal dragons stop censoring you and write some more. Apply chemotherapy of sorts to your thoughts, douse them with alcohol if necessary and purge them from your system, kick those surly words out from within, fling them away like the wretched ugly obstacles they are, sentences that still sit in your mind, blocking the sentences behind. All hail NaNo!!

  16. I think the biggest decision to make comes down to whether you want to write about computing or not. On the one hand, you have vast experience in computing, and writing usually flows more easily when you write in your area of expertise. On the one hand, after coding all day, you might not want to write about computing in this mythical “free time” you have acquired.

    If you decide to write about computing, I’d suggest avoiding the usual topics that anyone can and does write about (“how to get started with kernel hacking”, “introducing a new user to Linux”, “women in computing”, etc), barring some incredible new insight on one of those topics. Several people have already suggested in-depth filesystem stuff, which certainly seems useful, and not widely explored. I personally really like this suggestion from a previous comment: “People obsessing about problems I hadn’t ever heard of.”. Finally, you might seriously consider writing papers for publication in conferences and journals, or even writing a book.

    If you decide to write about something other than computing, that gives you many more options. The next big decision: fiction or non-fiction? Fiction could prove fun, though it has a much higher potential for writer’s block, either in the middle of writing something large, or when trying to come up with the next idea for something small. Non-fiction seems like it has more potential to allow you to write about something unique.

    A few random non-fiction ideas, drawing somewhat from familiar territory:

    • Stories about hacking on something not related to software, particularly if you have a creative solution.
    • Ideas for technical solutions to problems considered non-technical; ideally, problems that people haven’t written about yet, or for which you have a new idea. Past examples (which I wouldn’t recommend unless you have some stunning new idea) include monetary systems, voting systems, etc.
    • Non-technical aspects to technical problems. Consider Paul Graham’s essays here.

    Finally, if you want to keep yourself writing by starting a writing project which entails a regular commitment, and you have any artistic inclinations, you could always start a webcomic. (Or even if you don’t, you could collaborate on one; the webcomics community has no shortage of people who can draw but can’t write well.)

  17. A pleasant and unexpected echo of comp.arch past…

    Though I may vaguely remember that you were involved with Linux even back then.

    The only SF of yours that came up with a quick search was at fanfiction.net. Your affection for the social relevance in Between the Ages paralleling that in LOTR is understandable. One Way of Everyone seemed not quite as polished but pleasingly unpredictable (especially after you had set up such a seemingly predictable outcome).

    To get back on topic, Valerie’s prose is sufficiently disciplined, articulate, and comfortable (it could probably also be starkly uncomfortable if the circumstances called for it) that it would be enjoyable to see it turned to fiction, and her apparent SF role models are excellent (at least Vernor Vinge and Neal Stephenson – I’m about to take her up on her invitation to become acquainted with Charles Stross, in part because a quote from him in one of her old blogs reminds me of a prank at DEC three decades ago). I’ve respected her contributions to lwn for years but only just now became aware of her more extended on-line presence (and thank her for creating a location here where old acquaintances may unexpectedly meet, despite not usually being the blogging type).

    – bill

  18. I would love to read about the interesting parallels between kernel hacking and looking after cute baby goats. But maybe that’s for your memoirs, after you retire at 38.

  19. I’d like to read your writing on…

    * Your workstyle, scheduling habits, and approaches to new problems and fields of endeavor. Is there a characteristic sequence you follow? How do you balance reading papers, going to conferences, training and advising colleagues, hard-core hacking, writing documentation, and so on?

    * People who have influenced your work or inspired you.

    * Anecdotes of times you solved a problem cleverly, or watched someone do so.

    * Entertainment you enjoy.

    -Sumana Harihareswara

Comments are closed.