Follow Slashdot stories on Twitter


Forgot your password?

Comment: Re:not the point (Score 1) 328

by DrXym (#48930401) Attached to: Why Screen Lockers On X11 Cannot Be Secure
X is filled with APIs and functionality that no modern desktop has used in years. It requires numerous extensions to support a modern desktop experience but with serious caveats (e.g. compositor's extra latency and issues translating coordinate systems). Every app and widget set avoids X as much as possible by using middleware libraries to avoid this brain damage. Every app is pushing pixmaps around for the most part. Network performance is crippled by the amount of stuff being pushed and the amount of bidirectional messaging that goes into supporting. It has a woeful security model.

It may be in use but doesn't stop it being obsolete. Fortunately most dists will flip the switch and use wayland over the next year or two. And not before time.

Comment: Re:not the point (Score 1) 328

by DrXym (#48925835) Attached to: Why Screen Lockers On X11 Cannot Be Secure
"Good luck ever actually getting rid of it, considering it is what every *nix gui app runs on. Even if the switch to Wayland happens, most people will still be stuck with using XWayland constantly for a decade."

Virtually every *nix app runs over abstraction layers such as QT, GTK, Pango, Cairo etc. Assuming there are wayland backends for these layers then porting isn't as hard as you think. There may be vestigal bits of X to clean up and some edge cases that need more effort (screengrabbers, video players, browser plugins etc.) but porting the majority of apps will just port over. Aside from that, if you *did* have some ancient X app you could still fire up X over wayland just for that.

X will probably stick around as a core component for a few more years in most dists and then it'll be pushed off to the side as an optional package, available for those who want it but not installed otherwise because it won't be needed.

Comment: Re:not the point (Score 1) 328

by DrXym (#48925773) Attached to: Why Screen Lockers On X11 Cannot Be Secure
"the solution is to merely add an extra function call to the X11 API rather than rewriting the whole thing. Problem solved, if there is one."

X11 is an arcane and largely obsolete framework. The fact it needs so many damned extensions to be any way functional is precisely the reason that developers are keen to get rid of it. It's not secure, it's filled with arcane and obsolete code and it's terribly inefficient both locally and remotely. Fortunately it'll be moved aside and replaced by wayland over the next few years.

Comment: Re:First Sale (Score 1) 457

Yeah, well, EB games should be sued, if they don't have that warning printed in every store in large print. As well as Amazon and thousands of others.

EB games don't trade in digital games. They trade in physical media. And unless there is a registration code in the box then it's implicitly transferable.

Comment: Re:First Sale (Score 1) 457

That is not true: you are NOT "buying a license to a game", you're buying the game.

Sorry it is true and wishful thinking doesn't change it. Virtually all commercial software is covered by a EULA - end user licence agreement. You are buying a licence to use the software under the terms described by the EULA. If you run afoul of the terms then your right to use the software may be void or other penalties may apply.

In this particular case I suspect UPlay, Origin, Steam are reasoning that the licence is non transferrable, and since it WAS transferred from one person to another it has become void. That sucks but it's well known that they do this and if you want to avoid it, don't buy licences off some reseller.

Comment: Re:grandmother reference (Score 1) 457

This is no different from what happens on Steam all the time. I remember buying Left 4 Dead in a store for less than it cost to buy it on steam. The retail copy contained a steam code so it was effectively the same game.

It just demonstrates the utterly obscene pricing models in these online stores. In the real world the MSRP / RRP is just a guideline - the store can sell a game for any margin they like and usually they reduce it below MSRP. In the online store, the price is always the MSRP. I occasionally read the (pathetic) excuse that it's the publishers who set the price and there is nothing the store can do about it. Wrong! Publishers should be required to sell their digital download licences at the same wholesale cost as the physical copy and then digital stores retail can compete on their margins.

Just recently Sony offered a 10% discount off of PSN by way of apology for being attacked on Christmas day. The irony is that even with 10% off the prices there were still more expensive than a physical copy with the cost of middleman and postage thrown in. It's not just them of course - XBL is the same. And Steam. And Origin. And UPlay. They only time these services offer value is for games so old that their retail sales have flatlined and where people might pay $10 for a game in digital form that they wouldn't even bother with in physical.

Comment: Re:grandmother reference (Score 1) 457

What Ubisoft are doing is no different from what EA did recently or what Steam did before them.

Personally I think they should let people use these keys but the keys should unlockversions of the game that are heavily localized, thus negating any "advantage" people think they got from buying them. e.g. I bet Far Cry 4 and AC 4 are a lot less fun if the audio, text and subtitles are hardcoded to Thai and multiplayer to Thai servers.

As it is, I wouldn't be surprised if the terms of service allow them to do precisely what they did but I think there are better ways to discourage code selling.

Comment: Re:Modula-3 FTW! (Score 1) 484

by DrXym (#48912271) Attached to: Ask Slashdot: Is Pascal Underrated?
I think it's right to say vanilla Pascal was not a good language and so every implementation went off and implemented its own extensions, hacks, workarounds. Turbo Pascal, Free Pascal, Delphi etc. I was reading Gnu Pascal's features yesterday and it's amusing to read the "features" which are features cherry picked from other implementations. There wasn't any standard to maintain cohesion or enable portability and the entire platform suffered from that.

It's not like C/C++ is a perfect language - it's a horrible language in some ways but it's also very powerful and quite portable with discipline. It also has standards by which to measure implementations by and it has tended to keep compilers pretty close together aside from a few extensions.

Comment: Re:Modula-3 FTW! (Score 1) 484

by DrXym (#48902919) Attached to: Ask Slashdot: Is Pascal Underrated?
The readability of C depends on who wrote the code. It's not hard to find code which uses abbreviated variables all on one line, no comments, bad formatting, monolithic, copious use of MACROS, pointer abuse and all the rest to make unreadable code. Indeed, the international obfuscated C contest shows how easy it is to write utterly meaningless code which somehow does something.

And C++ adds it's own layer of fun. Templates are the work of the devil - get an arg wrong e.g. miss a const or a (de)reference, and the compiler might throw a wall consisting of hundreds of errors back at you. Not intuitive at all and certain not easy to step debug.

Doesn't mean the answer is Pascal but C/C++ was never designed for readability and any that exists is by the grace of the person who wrote the code rather than inherent to the language. Other languages do try a lot harder to enforce readability in the file structure and in the code itself. Python would be most famous for it but even enforcing filename = classname, path = namespace as seen in Java / C# gives more structure than you get in C.

Comment: Re:Modula-3 FTW! (Score 1) 484

by DrXym (#48902803) Attached to: Ask Slashdot: Is Pascal Underrated?
I envy you having to use Pascal at University. We had to use modula-2 which is even more anally retentive when it comes to boilerplate.

Anyway, Pascal lost the language war because it wasn't low level enough (at the time) to compete with C and didn't offer any other advantages. By the time franken-Pascals like Delphi appeared to gain those features it was too late because Java filled the application end and C/C++ was still there for the other stuff.

Comment: Re:Full Vaccination Wouldn't Stop This (Score 1) 660

by DrXym (#48884207) Attached to: Should Disney Require Its Employees To Be Vaccinated?
And if 100% of people wore seatbelts it wouldn't prevent some people dying in car crashes. Does that mean the exercise is worthless? After all, if a seatbelt isn't 100% effective what's the point? Except of course even if it were only 50% effective that still represents many thousands of lives saved every year, not to mention many tens of thousands more who suffered less traumatic injuries.

And in the case of a communicable disease, it needs pathways to spread. Block enough pathways and it cannot spread. This is what herd immunity is. Even if a few % of people cannot be vaccinated they are still surrounded by enough people who are. It is no coincidence at all that when these outbreaks occur it is ALWAYS in areas where the vaccinate uptake is lower than required for herd immunity to be effective.

"Love your country but never trust its government." -- from a hand-painted road sign in central Pennsylvania