Microblog
in reply to @ 2016-11 15:11 UTCI think we all wish laws and licenses and contracts and legal things were straightforward. Unfortunately, our society seems to be getting worse in this regard. Ignoring it will not make it go away.
in reply to @ 2016-10 17:58 UTC(a) will best best served (ish) by AGPLv3. (b) seems like a separate sort of concern
in reply to @ 2016-10 13:38 UTCRequestTracker (RT)
in reply to @ 2016-9 19:55 UTCWhat tor version are you running? Maybe it's a bug in my version.
Does it reconnect quickly when you come out of hibernate, or does it take some time?
in reply to @ 2016-9 04:05 UTCIf I connect to hidden services of various kinds, much of this advice should not matter, correct? Because there is no exit node so my circuit is not dirty?
in reply to @ 2016-8 22:36 UTCCompletely agree. Maybe we should make good microsites for some of the top stuff (like libreoffice)?
in reply to @ 2016-8 22:32 UTCIf you ever get stuck with a SomeException, you can downcast using http://hackage.haskell.org/package/base-4.8.1.0/docs/Control-Exception.html#v:fromException
in reply to @ 2016-8 22:30 UTCYeah. Telegram is pretty worthless
in reply to @ 2016-8 13:43 UTCIn this case all exceptions in the IO action have been excised, but readFile sticks a pseudo-exception into the pure value inside the IO action.
in reply to @ 2016-8 13:38 UTCA library to wrap the standard IO actions unexceptionally is something I've considered also.
There shouldn't be any other kinds of pseudo-exceptions that "should not be caught" since any other exception type is thrown on purpose. If you don't want to handle it at the call site, use ExceptT
in reply to @ 2016-8 04:37 UTCRankNTypes is by far the most expressive addition not yet in Haskell, and many things simply cannot be written (or must be written very differently) without it
in reply to @ 2016-7 14:40 UTCHaving watched your videos, you clearly "get it" when it comes to some of the issues the eternal copyright term extensions and DRM bring to our world, and especially the Internet. Have you considered fixing that for your own content by intentionally putting a libre license on things that have been out for a long time?
in reply to @ 2016-6 14:32 UTCDo you reject the idea that RYF-certified devices respect freedom? Independent of anything else they may also do?
in reply to @ 2016-6 13:31 UTCGiven that RYF means "endorsed by FSF" then of course you can't use terms or anything else they dislike. That's how you get advertising endorsements in any case, by submitting to the will of the endorser.
in reply to @ 2016-6 13:17 UTCThe OS is not the hard part. Drivers are the hard part
in reply to @ 2016-6 02:09 UTCOnce you're dealing with the copyright hold the CC license no longer matters at all. If you buy a different license then you abide by the terms of that different license.
in reply to @ 2016-5 21:20 UTCLaziness is a better default. You can always be explicitly strict if you need to
in reply to @ 2016-5 21:00 UTCIf you can convince them to go for a freedom-respecting license, that would be idea. But if you cannot, then I agree you are stuck with regards to proprietarising this particular libre sound effect. Perhaps you can buy one from a stock sound company or similar?
in reply to @ 2016-5 18:03 UTCCorrect. Unless you can make some sort of legal argument why you are not required to abide by the term of the copyright holder, you either cannot use their work at all, or must abide by their terms (which would include licensing your work as CC-BY-SA or one of the compatible licenses, such as ArtLibre).
The video itself is a strict safety guide so it cannot really be remixed or transformed.
Anything can be remixed. For example, a translation is a remix that seems likely to be useful in such a case.
in reply to @ 2016-5 15:31 UTCmutt is what I use daily, so let me know if you have questions ๐
in reply to @ 2016-5 02:01 UTCIf Thunderbird is too heavy for you then Claws.
Unless you like terminal and/or awesome things, in which case try mutt
in reply to @ 2016-4 20:04 UTCCould use ethercalc if a calendaring solution won't work, maybe?
in reply to @ 2016-3 19:29 UTC
in reply to @ 2016-3 13:10 UTClooking at a function signature you do not know if an exception can be thrown
Which is, in practise, the case in GHC. Sure, it's all well and good to say "IO can always throw an exception", but in practise you do not expect this of many IO actions.
in reply to @ 2016-3 13:02 UTCThat's why I added several more comments ๐
in reply to @ 2016-3 04:02 UTCOf course. As they copyright holder, you can grant anyone any other license you like.
in reply to @ 2016-2 23:25 UTCWhile you may choose to do this of course (it's your work, you may choose whatever you wish) — I would suggest looking into the SA clause and copyleft generally. This is a much bigger ask than just "a small attribution" and both much more beneficial to society and tends to discourage the sort of big-money corporate exploitation you probably fear most.
in reply to @ 2016-2 21:54 UTCFor a while I put syncIO from the errors package everywhere
Yeah. If you want to be extra paranoid then syncIO from https://hackage.haskell.org/package/unexceptionalio-0.2.0/docs/UnexceptionalIO.html is your friend. Forces you to never forget it anywhere. Then handle anything reasonable at the call site and propagate anything that needs to be with a differentiated type.
People like to hate on the typed exceptions in Java, but the issue there is the lack of type inference really.
None of this will help you with async exceptions or programmer errors, but those need to be handled in a different way anyway.