« first day (1935 days earlier)      last day (1245 days later) » 

9:06 PM
Just caught up on chat. Was thinking, if we assign IDs to inspections, would it be better to go MS-style and have an alpha prefix? RD0001 would be much more googleable...
 
IDK if it'll be RD001
IINM, MS' convention is to use 3 prefix for components
e.g. msbuild errors are MSB (?), C# errors are CS, and so forth
 
hmm good point - I was thinking of code analysis, docs.microsoft.com/en-us/visualstudio/code-quality/…
but yeah, I just meant something searchable
 
Please let us not start with something like this.
I do not want to have to look up ids when people open issues.
 
oh ok, was just a random thought :-)
 
9:13 PM
makes me wonder why MS do it anyway
I assume because they need a short and easy identifier
since they have many possible errors
 
It does make it more googable.
Instead of just searching the full text.
 
I suspect its the searchabilty thing. But they have a slightly wider user base...
 
Especially since they localize their error messages.
So this way, the code stays the same, the error message doesn't.
 
Our annotation names are kind of like ids, but readable ones.
At least if you speak English.
 
True. Resharper does the same.
 
9:16 PM
TBH, localization is the best argument for providing IDs too, IMO.
 
@M.Doerner aren't you worried about people conflating some inspection due to similar names... say, MissingAnnotationArgument vs. MissingMemberArgument
 
That makes me wonder, when will we see the first programming language in Chinese?
 
they mentioned that in the keynote today
> Uses MIT or other compatible license, and third party contributions are documented in a notice file.
Well, that's us out :-)
 
GPL doesn't fall under "other compatible"?
seems harsh
 
9:22 PM
Hmm, I assumed not, I assumed they meant "permissive"
 
It makes sense, this is MS after all.
 
But I could be wrong
 
Why would they promote writing software that cannot incorporated into closed source comercial software?
 
Exactly.
I'm beginning to hate the GPL license.
 
9:23 PM
I think it was of it's time
 
If I'm making something OSS, someone else should be able to use my work as a subset of their own without making their own completely open.
 
but they can?
the conditions only apply if they share it further...
 
@Hosch250 if you're making a library, yes
 
I'm not opposed to other people using it, but I only use Apache 2.0 (basically MIT, with a patent clause).
@MathieuGuindon No, it's more.
 
9:24 PM
Alternatively, you can use LGPLv3, which would enable them to use different licenses while using your component.
 
@Hosch250 or BSD
 
For example, assuming I found the time, I could build a source control feature for VBE.
 
@Hosch250 SC didn't need any of RD anyway (beyond the docking)
 
But it would have to be GPL because I've seen the safe COM wrappers we built and would likely have to reference them at some point.
@MathieuGuindon It needs the safe COM wrappers.
 
9:25 PM
maybe in future, RD'll be able to split and use that as a "Add-In base"
but that's wayyy out in the future
probably v5 or v6
 
fwiw I'm very much ok with the idea that no corp can fork RD and then enhance it behind closed doors, keep the enhancements to themselves, and the make money off our backs.
 
Thing is, even if we split out the GPL bits (SI and grammar, anything else?) how could we relicense? would that need agreement from every contributer?
 
Yep.
 
@mansellan welcome to hell
 
lol
 
9:28 PM
@MathieuGuindon I think you mispelt "Stallman"
 
You cannot split it out under a more permissive license anyway, if it needs to reference anything in RD.
 
@this R.M.S. TITANIC?
 
GPL is Hotel California. You can check out any time you like, but you can never leave...
A bit like the European Union, as it turns out!
 
and it was all because of a .g4 file that has since been fully rewritten
@mansellan LOL!
 
@mansellan I think you meant a Roach Motel.
 
9:31 PM
lol that too
 
(is there any difference between the two anyway?)
 
@MathieuGuindon wait, so our grammar would not need to be GPL today?
 
Alright, TTQW and go install Linux.
@mansellan awgaya rewrote it.
With a little bit of help later from others.
 
Just wondering then, if we do client/server, could the server be MIT if we wanted?
 
do a diff between the ANTLR's version and ours.
(unless ANTLR took ours since? I think yes because Scott seem to mention doing that, I think?)
 
9:34 PM
I'm not advocating anyone leave GPL. It's just that it's my opinion that if someone wanted to fork RD, modify it, and sell it, for example, they'd have to make some significant changes we could never do on our own, and anyone who needed those changes would be able to buy it.
 
Not, if the server uses the grammar or the indenter.
 
@this IIRC it was the v1.4.3 grammar that I PR'd back into the antlr grammars repo
 
TBH, I don't think RD should ever re-license as non-GPL anyway.
 
And we could keep up this one, and anyone who didn't want to buy it and didn't need those big features could use the free version.
 
sounds like a non-starter
oh well
 
9:35 PM
Let's sface it - there's too much commercial value in RD.
 
true
 
IMO shoving RD features behind a paywall destroys the very idea of RD
 
It would be a immense bummer if we find out that Acme corporation made their Rubberdick program
 
I just hate the one-way nature of GPL. Feels like a trap...
 
and everyone went over to it because it's better maintained.
 
9:36 PM
yep, that would suck
 
@this that typo...
 
Well, so is CSS for which GPL was designed to oppose.
@MathieuGuindon what typo? I don't see any typo.
 
you ever noticed the ironic parallelism between OSS and CSS and the bathrooms in gas stations?
 
I suppose things like Net Core can avoid the commercialisation trap of permissive licenses because they have huge resources, its unlikely anyone could outpace them enough.
 
9:38 PM
wait isn't CSS something something style sheet?
 
@mansellan .NET Core is, IIRC, Apache 2.0.
 
If anyone has ever went to those old-timey square service-center-turned-gas-station, they usually have bathrooms around the back and they lock it up and insist that you come in, take a key attached to a toilet seat or brick
 
oh I'm such a dummy
 
@Hosch250 yep, that's permissve?
 
but those bathrooms are absolutely filthy
 
9:39 PM
Basically MIT.
 
with a patent waiver
 
in contrast, those that allow free use of bathroom, usually have much cleaner and better bathroom.
 
Yeah.
 
I meant, there's no danger someone could take net core and do what @this postulated, because it's under such heavy and ongoing dev
with a smaller, commerially valuable project, it's more likely
 
Sure they could.
R# could.
RedHat could.
Oracle could (but avoids everything .NET).
Amazon or Google could.
 
9:41 PM
TTGH
 
But they'd have to add stuff people really wanted, and were willing to pay cash money for
and given how functional it is already, that could be tough
 
there's also the stack
it's too low on the stack to be swappable
you'd end up play catch-up forever
 
@this like if JetBrains got into VBE Extensibility and made a R# for the VBE? They've had 20 years to do that and haven't =)
 
TBH, if Matt hadn't OSS'ed RD, he could be working on it full-time now paid.
(It might've never gotten this far, though :P)
 
@Hosch250 eh? there has been companies that made a living on OSS, no?
 
9:44 PM
Mostly support.
 
yep
 
And RD doesn't need that much support.
 
I'd like to think that if we do ever do a add-in system though, that add-ins can be licensed however the author wants.
 
I meant, if it was closed-source, he'd probably be able to sell it like Wayne sells his stuff.
 
@MathieuGuindon Yeah, I guess they didn't think there was much commercial value for it.
 
9:45 PM
I was also contemplating a SC add-in, but would only do it as permissive
 
Sure, but it also means he'd be cowboying it all on his lonesome.
 
@Hosch250 and it wouldn't have had half the traction, userbase, and most importantly contributors. i.e. it would be just my work, and it would suck big time because I code like shit and never write enough tests for anything.
 
rofl!
 
It would be cool if there was a license that allowed works to be closed-source for a fixed period.
For example, if I decided I wanted to make a living working on RD.
 
hmm. Sun did have that model.
 
9:47 PM
And RD had this license.
I could keep my features for 1-3 years or whatever closed-source.
 
not sure if Oracle kept that.
 
I'd have to recoup my costs by then, and then I'd have to OSS each feature.
 
Well, presumably every license is ultimately secured by copyright, so will eventually go public domain once abandoned. that's something like author's lifetime + 70 years iiuc.
 
(and increased every time Disney needs to keep Mickey Mouse out of reach)
2
 
9:49 PM
(IIRC)
 
uh, if VBA's still in production 100+ years later.... dang
 
@this VBA will still be in production at the heat-death of the universe...
 
@mansellan Does public domain == open source?
 
more like unlicensable source -- see below
 
public domain means no copyright is possible
anyone can use any way they wish
and nobody can restrict it
MIT is more-or-less PD, just with a "don't sue me" clause, and with an attached ownership claim.
(the latter is needed to give basis for the former iiuc)
I think copyright is too lenghty across the board. 25 years from creation should be sufficient for more-or-less everything IMO, even high R&D like pharma.
oh wait, that's patents. meh, my point stands.
 
@this Interesting article. I've never really been on board with the idea of "owning" sequences of zeroes and ones, but pragmatically there needs to be some protection in order to reward invention. I just think the stupid IP lengths massively unbalance things, benefitting intermediaries to the detriment of society.
Hence, would be quite happy if it was 25 years across the board
</soapbox>
 
> don't write OSS for money. Have a full time job and do OSS on the side.
 
@MathieuGuindon Oh cool, I'm doing it right :-)
 
10:37 PM
Visual Studio 16.3 finally kills off the Start Page. Trying to see if I can resurrect it...
 
 
1 hour later…
11:45 PM
:gasp: Learned me somethin' new today: Canada's got rednecks too. Would have never guessed they'd be Canucks.
 

« first day (1935 days earlier)      last day (1245 days later) »