HNHacker News
TopNewBestAskShowJobs

Bill_Dimm

616 karma · joined June 11, 2012

The founder and CEO of Hot Neuron LLC. He developed the algorithms for conceptual clustering, near-dupe detection, and predictive coding used in the company's Clustify software. He has experience in the development and application of mathematical models to solve real-world problems in theoretical physics, mathematical finance, information retrieval, and e-discovery.
submissionscomments
Bill_Dimm··on Improved default settings for Linux machines
If you plan to hibernate, I believe you need at least 16GB of swap (i.e. the amount of memory you have).
Bill_Dimm··on The Browser Hacker's Handbook
Google cache:

http://webcache.googleusercontent.com/search?q=cache:au9GY9v...

Bill_Dimm··on Why the New Obamacare Website Is Going to Work This Time
Recording/streaming the event makes it very easy to take an out-of-context soundbite...

I would have concluded the exact opposite. Making a full recording available allows any attempt to quote part the meeting out of context discoverable and provable. Without a recording, anyone that attends the meeting could jot down a quote that serves their purpose and report it without anyone being able to dispute the context (aside from relying on the memories of other attendees).

Bill_Dimm··on Kivo: Simple document feedback in your browser
Interesting. Some comments:

1) It doesn't mention what file types it accepts. Since it says "presentations" I'm tempted to think .PPT, but some tools for presentations (like SpeakerDeck.com) take .PDF.

2) Why presentations? Why not documents (e.g. word processing files)? I haven't used Google Docs, but if it doesn't have feedback capability it seems like documents might be a more fruitful niche than presentations.

3) Perhaps related to (2), would it make more sense to highlight the text you want to comment on rather than drawing a box around it (which doesn't work so well if the text wraps to the next line)?

Bill_Dimm··on How we nearly lost our domain, and how to prevent this
Did they give an explanation for the contact info getting changed (assuming you complained about it)? I'm wondering if it might be related to a botched data transfer between them and Enom (which does the actual registrations for NameCheap). The Enom link does seem to create problems sometimes. I once tried to renew a batch of domains and it reported that a few failed, so I did those ones again and got double-billed for them (customer service fixed it promptly). Other than that, I haven't had any problems with them (no contact details changed, although I just remembered that their email to confirm contact info once gave wrong data [I think my apartment number was dropped or something like that] when the actual whois was fine).
Bill_Dimm··on China bans use of Windows 8 on government computers
They had one called Red Flag Linux, but apparently it shut down recently:

http://en.wikipedia.org/wiki/Red_Flag_Linux

Bill_Dimm··on The Curse of Dimensionality in Classification
Cross-validation (actually, this is mentioned toward the end of the article). Basically, fit the the classifier with a subset of the data and test the predictions on the remainder. Predictions for out-of-sample data will be poor if you have overfitting.

http://en.wikipedia.org/wiki/Cross-validation_%28statistics%...

Bill_Dimm··on Ask HN Mods: Can you make submission URLs human-readable?
I find long URLs to be annoying in email since you never know if they will get wrapped and broken somewhere along the way. And, if you're emailing a URL to someone you are presumably providing some text to explain what it is, so why do you need extra junk in the URL?

You can already add whatever you want to the URL if you really think it needs to be verbose:

https://news.ycombinator.com/item?id=7536719&title=whatever-...

Bill_Dimm··on Ask HN Mods: Can you make submission URLs human-readable?
Search engine spiders, scrapers, etc. would end up hitting a bunch of dupes when the title change causes a new URL to be generated that really points to the same content (not that spiders don't already have to deal with that mess on other websites).
Bill_Dimm··on New Linux version will reduce suspend and resume times
And, none of that is from the post you replied to; it is from my first post two levels up. The post you replied to is the one where I pointed out that my claim that it feels "sloppy" is based not merely on a few failed installs but also on a whole lot of bugs that have nothing to do with the hardware. Why you would reply to that and imply that it was my fault for not researching the hardware I was purchasing is a mystery to me. Your post would have made some sense (if you ignore the fact that I already stated that the hardware was certified for Ubuntu) in response to my first post, but seems really odd where you chose to put it.

> I can't see anything that you mentioned what wasn't hardware related.

I listed five bugs that have nothing to do with hardware, and you didn't address any of them. I really think you replied to the wrong post. https://news.ycombinator.com/item?id=7458973

Bill_Dimm··on New Linux version will reduce suspend and resume times
I tried Ubuntu several years ago, and really didn't think it lived up to all of the hype. I encountered a lot of obvious bugs (perhaps because I was using the 64-bit version, which was probably somewhat unusual at the time) and the admin tools seemed weak. Add to that some of the weird stuff that Canonical has been doing lately (privacy issues around their desktop search, Unity, Mir), and I'm just not too eager to go in the Ubuntu direction. If I hadn't gotten OpenSUSE 12.3 (mostly) working, Ubuntu would have been my next step.
Bill_Dimm··on New Linux version will reduce suspend and resume times
When I'm buying a laptop, I keep in mind that I plan to install Linux on it, and plan accordingly.

You do realize that the laundry list of flaky behavior described in the post are replying to has nothing to do with the hardware (except maybe the shutdown issues), right?

Also, as I stated in my original post this laptop was certified for Ubuntu (I found that out before purchasing), so I did keep in mind that I was going to install Linux on it when purchasing, and I was expecting drivers to not be a problem. You can call me a fool for expecting OpenSUSE to work on it when it was certified for Ubuntu, but there is only so much you can confirm in a advance (and, Ubuntu-derived Mint failed on it).

OS X wouldn't have installed at all, and the wrong version of Windows may have failed even harder.

This misses my point. I realize that Linux manages to accomplish quite a lot without the support from hardware vendors that Windows enjoys. But, between 1995 and 2007 there was a huge amount of improvement in "just works" and polish for Linux. Much to my disappointment, I don't think there has been that kind of improvement between 2007 and 2014. As I said from the beginning, my experience may just be a fluke. Or, maybe it isn't. Maybe it's caused by Novell no longer being behind OpenSUSE. Maybe it is caused by resources being spread across too many distros so none of them get enough polish. Maybe it is caused by a focus on selling enterprise licenses creating a disincentive to fix bugs in free versions. Or, maybe it's all just in my head.

Bill_Dimm··on New Linux version will reduce suspend and resume times
Well, I was installing pretty common distros on a pretty common laptop, and most people would have probably given up after expending many fewer hours than I did, so I think the risk of a random person trying to install Linux and coming away with the impression that Linux doesn't work is pretty high.

Also, it's not just install problems, and not just the problems that I listed in my original post. It's the fact that desktop search doesn't seem to be able to search the contents of my LibreOffice files when it worked just fine in OpenSUSE 10.2 (I've now taken to using "deepgrep" from the command line when I need to search my spreadsheets). It's the fact that task icons on the KDE panel sometimes end up on top of each other. It's the fact that when I leave the computer idle for a little while it tells me that it is locked and a password must be entered, but simply touching any key gets me right in (making the user think the computer is locked and it is safe to walk away from it when it isn't - nasty from a security standpoint). It's the fact that when I ssh into another computer and hit CTRL-D to log out ssh gets stuck and I have to kill it. It's the fact that it sometimes gets stuck when shutting down, so I have to hold the power button down to shut it off (which makes me worry about the possibility of corrupting something). Like I said, it feels sloppy.

Bill_Dimm··on New Linux version will reduce suspend and resume times
While clearly anecdotal, my own experience is that things have really not improved all that much over the past seven years. I've been using Linux since 1995, so I remember the bad old days, and things had improved a lot leading up to 2007 when I installed OpenSUSE 10.2 on a Dell D620 with relatively few problems. I recently tried installing Linux on a Dell E5430 (another mainstream business laptop, which is certified to work with Ubuntu) and my experience was not nearly as good:

OpenSUSE 13.1 was a disaster: https://bugzilla.novell.com/show_bug.cgi?id=835277#c31

Mint 16's installer had problems: http://forums.linuxmint.com/viewtopic.php?f=46&t=153379&p=80...

OpenSUSE 12.3 worked, but I had to mess around to get WiFi working: https://bugzilla.novell.com/show_bug.cgi?id=835277#c33 and the touchpad still has issues: https://bugzilla.novell.com/show_bug.cgi?id=869722 and there is other weird behavior, like the laptop screen not turning off when the laptop is in a docking station with the lid closed (did not have this problem with OpenSUSE 10.2), and the encrypted partition password prompt being off the screen if an external monitor is plugged in: https://bugzilla.novell.com/show_bug.cgi?id=862961

I'm a big fan of Linux, but things do not seem to have improved over the last seven years. It just feels sloppy and unstable. I'm sure some things have improved, but too much feels unpolished. Maybe I just have bad luck, but I was talking to a guy the other day who has been using Linux since 1996 who had the same impression (he was having problems with disk corruption on a desktop installation).

Bill_Dimm··on Why I killed my standing desk
Also, be aware that the muscles you use when running that way are very different (much more calf muscle involvement) and you need to build up to it gradually. I was told once to start off running 15% of the distance you did when heel-striking and build up gradually. Also, speaking from my personal experience, if you start to feel a calf muscle tighten up, stop IMMEDIATELY and stretch it. This may be helpful: http://www.barefootrunning.fas.harvard.edu/
Bill_Dimm··on HTML is almost 100% responsive out of the box
It responds well to being resized (e.g. to fit on a mobile phone screen instead of a large monitor). Although, I think it is normally used to mean considerably more than the simple scaling being done here (e.g. I've seen designs where the text goes from two columns to one if the display is narrow). http://en.wikipedia.org/wiki/Responsive_web_design
Bill_Dimm··on A competitor emailed me moments after I signed up to Stripe
That is creepy. I'm not sure whether the publicity from the blog post is going to generate more or fewer customers for the Stripe competitor, though.
Bill_Dimm··on The LeBron James of Programmers
For many years I have wondered why so many people think that supply versus demand ought to be a linear function.

Many curves are linear if you zoom in on a tiny enough piece of them. If there are a million 1x programmers in the world, and you increase the supply by a single 1x programmer, the price for a programmer drops, in theory, but only by a miniscule amount. Moving one part in a million along the curve is not enough for the non-linear terms to be significant compared to the linear term. Likewise, adding a single 10x programmer is still almost completely linear -- you are adding 10 more units of programming capacity to the million units that already exist. We aren't talking about a 50% change in supply here, we are talking about a 0.0001% change.

Now, if you are a tiny company that needs maybe two 1x programmers, you aren't going to pay 10x for a 10x programmer that will be idle for 80% of the time, but why wouldn't a large company that needs hundreds of programmers be willing to pay 10x for a 10x programmer instead of paying for ten 1x programmers when the cost is the same to achieve the same amount of work?

One final analogy: You walk into a store and find that a single roll of toilet paper costs $0.50. Would you expect a 20-pack to cost vastly less than $10 (say $1.00 or $2.00) due to supply and demand? There are practical reasons for there to be some discount for the larger package (less packaging per roll, less human labor to deliver to store shelves, etc.), but a discount beyond that?

By the way, the whole NBA analogy is terrible. The goal is to win the game, so being a little better (in terms of points scored) than the average player can have a big impact on the binary win/lose outcome. You don't need to score 1000x as many points to be worth 1000x as much money. In sharp contrast, much of software development is not so binary (perhaps just a few highly technical areas where an average programmer may never create an algorithm that is good enough to work).

Bill_Dimm··on Tesla Beats In Q4 With Adjusted Revenues Of $761M
Yes. It's short for "Tesla's earnings beat expectations in Q4"
Bill_Dimm··on Netflix performance on Verizon and Comcast has been dropping for months
Interesting, but what percentage of people have Apple TV? Is it enough to have that much of an impact?

Note: Edited to remove part that was actually addressed in original post.

Bill_Dimm··on Tex – Show off your best scientific illustration
If you have any text labels in the diagram, the font will be consistent with the rest of the text.
Bill_Dimm··on Graphs Show How Reddit Got Huge by Going Mainstream
Also, the ones in this article are square shaped, so you can't even see total site growth over time in them, which would be interesting and relevant.

I came here to say that (although your other points are good, too). The title of the article is "Fascinating Graphs Show How Reddit Got Huge by Going Mainstream" but then they normalize away all of the growth info and make it completely impossible to see how traffic growth is related to anything they are displaying.

Bill_Dimm··on I Decided to Delete All My Facebook Activity
I don't know if this procedure for deleting a Facebook account still works, but it's worth a try: http://www.groovypost.com/howto/security/permanently-delete-...
Bill_Dimm··on New Study Finds That More Attractive Students Have Higher GPAs
Original article has much more detail: http://healthland.time.com/2013/12/10/teachers-give-better-g...
Bill_Dimm··on Stupid Things Finance People Say
The expression is normally used to comment on the overall market behavior for the day (or some other time frame), not a single transaction, so the whole "There's one buyer and one seller for every trade" thing is somewhat irrelevant (edit: what I'm trying to say here is that John shouldn't count as two people just because he is involved in two transactions since the statement applies to the aggregation of all transactions over a day or other time frame).

Also, from John's perspective it may look like a single trade (he won't know that a computer matched his sell with multiple buyers unless they buy at different prices).

Bill_Dimm··on Stupid Things Finance People Say
"More buyers than sellers."

This is the equivalent of saying someone has more mothers than fathers. There's one buyer and one seller for every trade. Every single one.

If John sells 1000 shares and Frank and Sam buy 500 shares each, how many buyers and how many sellers are there? Granted, that isn't the scenario envisioned by people that use the expression, but if you are going to write an article full of nitpicking...

Bill_Dimm··on “Remove occurrences of the short name of the Debian derivative from Canonical”
That really depends on whether or not I could afford to (the sad reality of our legal system). Fortunately, the EFF seems to be helping him: https://micahflee.com/2013/11/canonical-shouldnt-abuse-trade...
Bill_Dimm··on “Remove occurrences of the short name of the Debian derivative from Canonical”
I don't really see how Canonical can take away the right to use a trademark nominatively[1], as provided by law, just by declaring your rights to be gone in a policy.

[1] http://en.wikipedia.org/wiki/Fair_use_(U.S._trademark_law)

Bill_Dimm··on MacKenzie Bezos Writes Amazon Review for Jeff Bezos Biography
Yep. Several years ago the Eastern Technology Council (a local business association) had its annual party on the campus of one of its member companies named Safeguard Scientifics. The local paper reported it as a Safeguard Scientifics networking or recruiting event (I forget which, but no mention of ETC), and claimed that resumes were being exchanged (I was there and didn't see a single resume, just people from various companies chatting, eating, and exchanging business cards). The article was so off-target that people on the Yahoo finance board for Safeguard Scientifics were claiming that this was a sign that the company was in great shape because it was on a big hiring spree. Reality: The company allowed a party to take place on its property, that is all.
Bill_Dimm··on Ember Charts: An easily extensible chart suite built with D3.js and Ember.js
Thanks! It's unfortunate that it's Chrome-only. It worked well on their examples, but not on this (not a d3.js):

  http://www.cluster-text.com/tmp/news.svg
Still, it's an interesting start.

Correction: It somewhat works if you wrap the SVG in HTML:

  http://www.cluster-text.com/tmp/news.html
although, when I load it into Inkscape the positioning of some text labels is way off (some aren't too bad) and some hidden elements are displayed.
← PreviousPage 3 of 7Next →