Hacker News Reader

Comment by mevile | Hacker News Reader

In this comment thread there are many people who did not read the entire article to its conclusion, and then also failed to read the many comments about the very last and very important bit at the end of the article.

There ought to be like a "didn't read the article" link next to the comment people can click to quickly bury those comments. It would avoid people having to reply again and again to say the same thing that commentator could have just learned had they read the article. Especially, you know, this bit:

> Remember: There’s a reason we gave the %localappdata%\lxss\ folder ‘hidden’ & ‘system’ attributes

There are multiple comments in this thread right now with that line in them.

View on HN

>There ought to be like a "didn't read the article" link next to the comment people can click to quickly bury those comments.

There is, it's the downvote button. The distinction of why a downvote is useless.

What are the requirements to be able to down vote?

a karma threshold. 800 or something?

500, I only just gained downvote ability in the past few weeks, when I finally gained more than 500.

Edit: another thing I should have mentioned is that it seemed to me that only those comments made after I gained this ability were downvotable, the thread my last few comments were in was sprinkled with downvotable comments, which were newer, and non-downvotable comments which were there before mine. I might have been imagining this or be mistaken though.

As a data point, I am at _exactly_ 500 right now, and cannot downvote. So perhaps you have to exceed 500?

EDIT: checking again at 507 and I now have the power to downvote. So there's our answer (plus my moderately shameless play for some upvotes worked).

Yup. 501. Remember. With great power comes great responsibility.

Indeed - compared with the ability to speak freely, the ability to downvote is not much of a privilege.

Council of 300

https://xkcd.com/1224/

the real question is if it goes away if you're downvoted.

I believe it does, but don't have any evidence to back it up. If itp volunteers, we could test by down voting his comments :)

Exactly... This effect might reinforce 'upvote'-grabbing behavior, to either not lose the ability, or regain downvote if lost.

Upvote-grabbing behavior is already rampant here, just as with any forum with point tallies attached to comments. Proof: all of the "edit: why the downvotes?" comments.

it seemed to me that only those comments made after I gained this ability were downvotable

I think you might be mistaken. You might not have noticed yet that the downvote option disappears on its own a set time (24 hours, I think) after the comment was made. Perhaps it was only the comments older than this threshold that were non-votable?

That's quite possible. I had not realized that the downvote option disappeared so quickly.

I think my karma accrued overnight, so when I noticed I'd exceeded the threshold a lot of the thread was perhaps a day old. The only thing I am relatively certain of is that some of the older comments in the thread were not downvotable, and the newer ones were.

only those comments made after I gained this ability were downvotable

That would be quite computationally intensive to implement.

Not if you store the timestamp of when you gain downvote rights, and compare against that.

> The distinction of why a downvote is useless.

Actually I think it would be incredible useful.

You could use it to shift a community to a more useful direction.

A more direct way of eliminating common errors.

If on a cool new tech that's renewable if someone starts another bloody global warming discussion being able to downvote as OT could try and keep people on topic rather than having to down vote something you agree with.

There are issues with the idea, it can become censorshop, but I find it an interesting one.

They could just make flagging dual purpose. But flagged comments probably alert the admins. Maybe clicking the flag button could spawn a drop down but I think there might be an aversion to introducing complexity and impedance. Requiring multiple clicks and using drop down elements might be a departure from the deliberate simplicity.

Colorize according to votes. Purple = off-topic. Red = troll. Orange = weird and scary.

You could blend colors too. Reddish-purple = off-topic troll. Dark orange = weird and scary troll. etc.

>There is, it's the downvote button. The distinction of why a downvote is useless.

How about a tag system associated with each posting? Then you can have just a scalar "weight" and a set of keywords (string labels), each with a designated meaning. You press a "plus" button or something, and a menu with optional tagging pops up (tag clouds can be used).

Some tags could be "spam", others could be "didn't read" and users can build their own threshold filters for tags. (0) spam; (4) didn't read; (3) incomplete; ...

I've thought about this, too. In particular, I'd like to, say, separately up vote "substantive" and down vote "agree". That said, how complex do we really want the UX to be?

As for "didn't read", if you're not interested, just move on. If you think the article is inappropriate and can, flag and move on. If you see a comment that looks like the poster didn't read the article, you can call civilly call out what is in the article, and perhaps down vote the comment.

> how complex?

Yeah I thought about this very hard actually, and came to the conclusion that just a "weight"(+) button is enough.

And, if you care that much so that you would waste your time providing more ... contextual meta-info (about the weight itself), you simply need a pop-up and a quick way to select words from a common dictionary. Certain users can even make content "more spammish" by simply increasing the weight with a click (they agree with the labeling and they are "special" in some way).

This, and the weight itself should provide enough info for a comprehensive "decoration" of the postings.

So... slashdot?

Mmm, to be honest, I was never a slashdotter so I have no experience with it at all (strangely enough :))

It has/had exactly that system but with also positive descriptors.

I can't fault you - slashdot hasn't been relevant for.. at least a decade.

I don't see a downvote button. Do you need a base reputation for that, like Stack Overflow?

Edit: yeah, karma is needed.

Also, isn't the entire new Linux/Windows subsystem still in beta? IIRC it's only in the preview builds and even then you still need to enable it explicitly.

From initial glance, it seems the problem could easily be fixed with a filesystem filter driver ( https://msdn.microsoft.com/en-us/windows/hardware/drivers/if... ) detecting modifications to paths under the Linux root directory. Actually it could be done entirely in userspace, this definitely isn't some insurmountable architectural problem.

Lolling hard at suggestions to use Cygwin as a viable alternative to WSL because of this tiny nit.

Not as simple as this because some tools create a working copy of a file when it's being modified, and then replace original file once changes are saved. Mapping these changes to the same file "identity" likely v. error prone.

It's available in the current public release, you still must enable it explicitly like other optional Windows features (like Hyper-V)

> Remember: There’s a reason we gave the %localappdata%\lxss\ folder ‘hidden’ & ‘system’ attributes

So basically "there is a gaping hole in the middle of the street but it's ok, I've put a 'road ends here' sign in front of it so now it's the drivers' fault if something happens"

In this case this results in a lot of files not being editable by windows programs, plain and simple - such as the entirety of /etc and /home. That's a pretty bad situation.

The metaphor is wrong. As a normal user you would never come into contact with these files. The "gaping hole" is actually in a fenced-off area with keep-out signs where nobody would wander in by accident.

You say that, but I've seen a few too many people delete random system files because they didn't know what they were and needed space. Granted, they should've read the "keep out" signs and have only themselves to blame.

These files are hidden by default, so not only didn't they read the "keep out" signs, they tore them down and said "See? No sign." and then ran into the hole.

I guess at some point you can't protect people from themselves anymore. One might even argue that people who use WSL (especially in beta) would be tech-savvy enough to understand how it's implemented and works, but apparently not ...

Sure - it is not a good situation. To me it would have sounded perfectly reasonable to edit WSL files with a Windows editor. There could at least be some kind of __README_FIRST_BEFORE_DOING_ANYTHING_IN_THIS_DIRECTORY.TXT.

I just wanted to make clear that the parent comment is slightly hypberbolic.

But then, a "normal user" wouldn't use WSL. Developers and power users do.

I create a file in my WSL home directory, and want to edit it in Sublime. So I google "where is the WSL home directory stored" and get `%LOCALAPPDATA%\lxss`. I plug that into WIN+R, and get the files in explorer. At no point was I stopped or even warned.

The reason is clear. WSL is grafted on to Windows, and has low priority. The windows shell and filesystem teams are currently not working to improve integration from their side (at least not publicly visible).

The shell people could put up a banner in explorer, like you got in Program Files in previous windows versions, that you shouldn't edit these files (e.g. via desktop.ini). From the filesystem side, it should be possible to make the kernel know about WSL and prohibit writing to those files from windows.

WSL could also have been implemented with an image file, instead of individual files in `%LOCALAPPDATA%\lxss`. The reason they did it this way, I believe, was to keep open the possibility of future two-way-integration. But this would mean change-listeners not only in Linux land, but in Windows-land... which again would need deeper integration from other groups at MS. I suspect once WSL has proven itself a bit, they'll add full transparent two-way-integration, i.e. you'll be able to edit lxss files from windows one day.

"WSL is grafted on to Windows, and has low priority" - not sure the team of ~12+ of us would agree with you on that.

And how would the shell team put up a banner in Sublime?

If the kernel prevented these files from being written to, how would we install the distro in the first place?

Could, should, might, may. Yep, there are 100 ways in which some things could, and may yet, be done, but these choices were made for hundreds of sound reasons.

We're still working hard on WSL - lots of improvements in the pipe.

> So basically "there is a gaping hole in the middle of the street but it's ok, I've put a 'road ends here' sign in front of it so now it's the drivers' fault if something happens"

What you're describing there is road works and yes of course it would be the drivers fault if they drove into the middle of the work. Even aside the obvious point about the work being sign posted, it's the drivers job to be vigilant of hazards - including those that aren't signposted (eg potholes).

> In this case this results in a lot of files not being editable by windows programs, plain and simple - such as the entirety of /etc and /home. That's a pretty bad situation.

Think about it more like using the right tools to edit the right job. You wouldn't blame Windows / Microsoft if someone screwed up their registry because they edited that binary blob in Notepad instead of Regedit. Equally if someone is clued up enough about Linux to install and use the Linux subsystem then it's reasonable enough to expect them to use Linux tools to modify the files on the Linux filesystem.

Reply on news.ycombinator.com