back
user profile

kristjansson

4,853karma·1,566submissions·February 17, 2015
recent activity (1,566 total)
comment
This is a maximalist position and an admirable goal. As far as I can tell though, this forces Nix (or the developer) to get into the business of re-packaging entire ecosystems.
4y ago·view thread
comment
Convenient (version-addressed!) FHSUserEnv is exactly what I want out of Nix. Land me in an environment that has a list of deps (at specific versions, not hashes!), let me go mess with it.
4y ago·view thread
comment
Usually part of the definition is holding yourself out to not married
4y ago·view thread
comment
Yup, the programs could have been better implemented, and requiring action from the tenant (presumably to prevent landlords from claiming 'unpaid' rent on empty units) supposes more cooperat…
4y ago·view thread
comment
There were subsidies provided by multiple levels of government[1][2], down the city level where I live. Apparently there was at least (if not more than) the usual level of misappropriation of those s…
4y ago·view thread
comment
Once you have billions and billions in profits to reinvest there's hardly a choice, is there? At some point, the firm has to invest in new product lines to support or supplant its tentpoles, and …
4y ago·view thread
comment
Mostly no, since PostgreSQL 12
4y ago·view thread
comment
Years of unpicking others use of Rs sqldf (which by default used to copy the entire data frame to a SQLite db, run the query, the copy the result set back) when they complained their R code was to slo…
4y ago·view thread
comment
This is the biggest thing. Indexes are like an extension of array broadcasting to relational data. Not least in that both broadcasting and pandas indicies seem surprising and magical if one doesn’t u…
4y ago·view thread
comment
R’s waxing popularity is a testament to that power. dplyr and friends effectively reject much of the R-the-language and substitute their own friendlier, more popular syntax without rejecting R-the-pla…
4y ago·view thread
comment
I don’t get that part either. OTOH, an effective property tax rate of 3.2% or so seems much more reasonable that the 9.8% percent implied by his table. Assuming people but anywhere close to as much …
4y ago·view thread
comment
Tying property tax to market value is a form of this, I think. The town doesn’t realize substantially more costs when property values go up, but wants to participate in the growth of ‘its’ property. …
4y ago·view thread
comment
I think the point is that, like a lot of cost problems, a relatively small proportion of users end up accounting for a relatively large share of costs. Just excluding those expensive users (implicit…
4y ago·view thread
comment
One instance is evidence their process failed at least once. At a minimum they need to identify the process failure and the cars the could have been affected, and issue a recall or TSB to verify none…
4y ago·view thread
comment
I’ll await Teslas recall of all units possibly affected by this issue
4y ago·view thread
comment
Income over the next 4 quarters is, of course, the only factor by which one can distinguish occupations.
4y ago·view thread
comment
Using models from your ORM of choice instead of the DDL they generate isn't that unreasonable. Suggesting that one should (or then could only?) use 'a simple API .. or UX' to actually …
4y ago·view thread
comment
It seems obvious, but one would be surprised how often the argument has to be made. Though the real argument is that Tableau Workbooks, Alteryx ETL-graph-things, and the whole panoply of opaque propri…
4y ago·view thread
comment
Where's Uber without a trustworthy, city-corner-accurate GPS location for every driver and rider? Where's Instacart without a high-quality barcode scanner^W^Wcamera in the pocket of every s…
4y ago·view thread
comment
There's an advantage to long-lived interpreters/REPLs on remote machines for the kind of work done in notebooks. Significant amounts of data may have to be read, expensive computation perfo…
4y ago·view thread
comment
Avoid if possible, is the easiest answer. Encourage your colleagues to move their code into proper packages when they're happy with it, and restrict notebooks to _use_ of their code. Failing tha…
4y ago·view thread
comment
Many, many people choose to reason only in binary terms.
4y ago·view thread
comment
All code is craft; the systems they implement are engineering.
4y ago·view thread
comment
He had two things: (1) the right to make and publish a change to a project he controlled and (2) a mechanism (npm and its ecosystems versioning practices) to have his changes run on thousands of machi…
4y ago·view thread
comment
The developer owns a copyright to the specific work - the literal arrangement of characters in a specific order, and non literal aspects like structure or organization. Copyright is created when the …
4y ago·view thread
comment
NPM/MS/GitHub are distributing code consistent with the terms of the license provided to them. The developers (current lack of a) relationship with his (former) service provider doesn’t hav…
4y ago·view thread
comment
The problem is that, AIUI, each NPM package installs its own, isolated set of dependencies. It has facilities to manage versions down in the dep tree, but it’s not as simple as other managers which u…
4y ago·view thread
comment
Sure, but there are points in the design space between “everyone specifies exact version requirements” and “everyone specifies lower bounds and pulls the absolute latest”.
4y ago·view thread
comment
Sure, he can do that. But everything else is released under an open source license. Anyone that has a copy (like, I dunno, NPM) can continue to distribute whatever version of the software they want …
4y ago·view thread
comment
Sure, he’s free to make new versions of it that do whatever he wants - spam zalgo, mine bitcoin, fire the mussels, whatever. That doesn’t obligate anyone to _use_ that code, nor does it obligate NPM …
4y ago·view thread