Enough for V1.0?
Locked
 
|
Hi All,
The core Redline Smalltalk compiler is working and I have started to write Apps with Redline Smalltalk. These apps have been other tools for Redline, like Stout and Stir. Given this my question is, should we mark the code as version 1.0 and notify people we have reached this milestone? I'm also thinking that with the announcement of V1.0 we may get more interest and people trying Redline. We would then improve and build upon V1.0 and hopefully gain momentum. What do you think? Rgs, James. |
|
|
At the very least, a V1.0 RC ... with a feedback period before it's declared done. I'd like to get back into it myself, having been distracted recently by various personal & professional developments. It's been in the back of my mind and some kind of announcement is helpful to refocus attention to it. :)
- Bob
On Sat, Apr 28, 2012 at 11:51 PM, James Ladd <[hidden email]> wrote: Hi All, |
|
|
Yes I think a v1.0rc would be the way to go. How long should an rc last? Is there a criteria for accepting or rejecting an rc? For me the renewed focus will be ideal. Sent from my iPhone
|
|
|
On Sun, Apr 29, 2012 at 05:46:06PM +1000, James Ladd wrote:
> Yes I think a v1.0rc would be the way to go. > > How long should an rc last? > Is there a criteria for accepting or rejecting an rc? > > For me the renewed focus will be ideal. > > Sent from my iPhone > > On 29/04/2012, at 5:33 PM, Robert Calco <[hidden email]> wrote: > > > At the very least, a V1.0 RC ... with a feedback period before it's declared done. I'd like to get back into it myself, having been distracted recently by various personal & professional developments. It's been in the back of my mind and some kind of announcement is helpful to refocus attention to it. :) > > > > - Bob > > > > On Sat, Apr 28, 2012 at 11:51 PM, James Ladd <[hidden email]> wrote: > > Hi All, > > > > The core Redline Smalltalk compiler is working and I have started to write Apps with Redline Smalltalk. > > These apps have been other tools for Redline, like Stout and Stir. > > > > Given this my question is, should we mark the code as version 1.0 and notify people we have reached > > this milestone? > > > > I'm also thinking that with the announcement of V1.0 we may get more interest and people trying Redline. > > We would then improve and build upon V1.0 and hopefully gain momentum. > > > > What do you think? > > > > Rgs, James. > > > > > > I don't think the 1.0 is there at all. A modern software release isn't just about code maturity, it's also tool integration, documentation and most importantly community validation. James has been able to use Redline to build out new pieces of the Redline infrastructure but is it ready for people to start writing applications with it? The Getting Started with Redline Smalltalk article runs hello world, but it doesn't show us a way to run our own projects. I think it's also a little awkward to declare a Smalltalk 1.0 when much of the standard library is pending implementation. Are you talking about marking only the compiler as 1.0? Since your goal for this 1.0 is getting people interested and gaining momentum, I think our documentation story needs to be better. Also, you mentioned a star (smalltalk jar) format on Twitter a few days back, but I don't recall seeing anything on it here on the mailing list. I think that belongs in our 1.0 story and would like to see some development releases making use of stars before committing to a format. One last thing that would give us a really strong 1.0 would be a tool like leiningen for Redline Smalltalk. For those unfamiliar, leiningen is the Clojure build tool, one of the things that has made Clojure so strong and given it really good penetration is the community's emphasis that the clojure jar is just another library that you have on your system and you can embed it in existing java apps and start out writing strategic subsets in Clojure. Complementing this is leiningen which is essentially the command line interface to Clojure. If anyone is interested in writing this build tool with me (in Redline Smalltalk of course) I am going to be working on it during late May and June. Right now I have to focus on finals. :( I think Redline is awesome and I want to continue working on it but we only get one shot at our 1.0 release and if we want it to be as effective as possible in getting more people to try it and use it, I think we need: 1. Thorough documentation on what works and what is pending implementation. 2. A couple months community use of the star format. 3. A leiningen like tool which is capable of bootstrapping itself. Please don't take this as an attack on your hard work James, I think the progress is astounding! -- Steven! nuclearsandwich |
|
|
No offence taken - this is exactly the discussion I was looking for.
V1.0 is meant to be for the compiler only. I'm thinking that while some of the tooling should be there, won't the community want to contribute too? Sent from my iPhone On 29/04/2012, at 6:10 PM, Steven! <[hidden email]> wrote: > On Sun, Apr 29, 2012 at 05:46:06PM +1000, James Ladd wrote: >> Yes I think a v1.0rc would be the way to go. >> >> How long should an rc last? >> Is there a criteria for accepting or rejecting an rc? >> >> For me the renewed focus will be ideal. >> >> Sent from my iPhone >> >> On 29/04/2012, at 5:33 PM, Robert Calco <[hidden email]> wrote: >> >>> At the very least, a V1.0 RC ... with a feedback period before it's declared done. I'd like to get back into it myself, having been distracted recently by various personal & professional developments. It's been in the back of my mind and some kind of announcement is helpful to refocus attention to it. :) >>> >>> - Bob >>> >>> On Sat, Apr 28, 2012 at 11:51 PM, James Ladd <[hidden email]> wrote: >>> Hi All, >>> >>> The core Redline Smalltalk compiler is working and I have started to write Apps with Redline Smalltalk. >>> These apps have been other tools for Redline, like Stout and Stir. >>> >>> Given this my question is, should we mark the code as version 1.0 and notify people we have reached >>> this milestone? >>> >>> I'm also thinking that with the announcement of V1.0 we may get more interest and people trying Redline. >>> We would then improve and build upon V1.0 and hopefully gain momentum. >>> >>> What do you think? >>> >>> Rgs, James. >>> >>> >>> > > I don't think the 1.0 is there at all. A modern software release isn't > just about code maturity, it's also tool integration, documentation and > most importantly community validation. James has been able to use > Redline to build out new pieces of the Redline infrastructure but is it > ready for people to start writing applications with it? The Getting > Started with Redline Smalltalk article runs hello world, but it doesn't > show us a way to run our own projects. I think it's also a little > awkward to declare a Smalltalk 1.0 when much of the standard library is > pending implementation. Are you talking about marking only the compiler > as 1.0? > > Since your goal for this 1.0 is getting people interested and gaining > momentum, I think our documentation story needs to be better. > > Also, you mentioned a star (smalltalk jar) format on Twitter a few days > back, but I don't recall seeing anything on it here on the mailing list. > I think that belongs in our 1.0 story and would like to see some > development releases making use of stars before committing to a format. > > One last thing that would give us a really strong 1.0 would be a tool > like leiningen for Redline Smalltalk. For those unfamiliar, leiningen is > the Clojure build tool, one of the things that has made Clojure so > strong and given it really good penetration is the community's emphasis > that the clojure jar is just another library that you have on your > system and you can embed it in existing java apps and start out writing > strategic subsets in Clojure. Complementing this is leiningen which is > essentially the command line interface to Clojure. > > If anyone is interested in writing this build tool with me (in Redline > Smalltalk of course) I am going to be working on it during late May and > June. Right now I have to focus on finals. :( > > I think Redline is awesome and I want to continue working on it but we > only get one shot at our 1.0 release and if we want it to be as > effective as possible in getting more people to try it and use it, I > think we need: > > 1. Thorough documentation on what works and what is pending > implementation. > 2. A couple months community use of the star format. > 3. A leiningen like tool which is capable of bootstrapping itself. > > Please don't take this as an attack on your hard work James, I think the > progress is astounding! > > > -- > Steven! > nuclearsandwich |
|
|
I think a 1.0 should be ready for production-use. So we should define some milestones and a timeline. -- Am 29.04.2012 10:39 schrieb "James Ladd" <[hidden email]>:
No offence taken - this is exactly the discussion I was looking for. |
|
|
In reply to this post by James Ladd
On Sun, Apr 29, 2012 at 8:46 AM, James Ladd <[hidden email]> wrote:
It can be some arbitrary length depending on our confidence in its readiness. I'd say a few weeks at most, one week at least, and the criteria for sticking the fork in it and calling it done would be that no show-stopping defects remain on the anointed date.
So then the question is: What's a show stopper? I suppose we can be more liberal with that definition in the early going, and more stringent as it gets more mature.
If RC1 fails, then we lather, rinse, & repeat with RC2, ... RCn, as needed. In reality I don't think Redline is, as a whole, is 1.0-ready, but then, the RC process can help us clarify what the criteria for completion actually should be. My guess is the compiler is in 'good' shape but subtle changes to it would impact all the tests for the runtime -- possibly dramatically -- so it probably makes sense to version them separately.
- Bob
|
|
|
Minor correction. I said: "I suppose we can be more liberal with that definition in the early going, and more stringent as it gets more mature."
Perhaps because the coffee is still filtering through my bloodstream, I meant the opposite: "I suppose we can be more stringent with that definition in the early going, and more liberal as it gets more mature."
Now is the time to get particularly the compiler in a rock-solid state and that means being more exacting in our estimation of what is and is not acceptable, in terms of known issues. This may *seem* to delay our gratification in terms of hitting the v1 milestone, but it will make it much easier to hit future milestones with momentum.
- Bob On Sun, Apr 29, 2012 at 10:03 AM, Robert Calco <[hidden email]> wrote:
|
|
|
I think 1.0 indicates something that is there.
I want rock solid for that. I'd call it 1.0 alpha 1 When we shake out bugs there, improve class library etc
( probably several alphas ), it goes to beta etc. What goes into each, that is an open question. On Sun, Apr 29, 2012 at 5:07 AM, Robert Calco <[hidden email]> wrote:
|
|
|
In reply to this post by Stefan Krecher
What would be the benefit of implementing JSR 223 support? I don't think Clojure or Scala currently implement that? Smalltalk isn't a scripting language - we actually compile everything to bytecodes before we execute.
How do we define "production use"? The compiler runs today, simple scripts are running, James has some webapps running via Stout. I can see us doing a beta period (like, 30 days?) and a subsequent RC and fixing up blocking bugs but not introducing new features during that time. Maybe implementing more of the base library during the beta period? - Rob On Sunday, April 29, 2012 1:56:25 AM UTC-7, Stefan Krecher wrote:
|
|
|
I'd like to see class library solid before beta label.
1.0 as full version means something to people and if we don't live up to it, we probably won't get another chance. Super early adopters will run with it no matter what the version is.
1.0 needs to aimed in terms of what non early adopters would expect.
On Sun, Apr 29, 2012 at 1:32 PM, Rob Roland <[hidden email]> wrote: What would be the benefit of implementing JSR 223 support? I don't think Clojure or Scala currently implement that? Smalltalk isn't a scripting language - we actually compile everything to bytecodes before we execute. |
|
|
I'll put a roadmap together for comment Sent from my iPhone
|
|
|
In reply to this post by Robert Roland
Hi,
Am Sonntag, 29. April 2012 19:32:33 UTC+2 schrieb Rob Roland: What would be the benefit of implementing JSR 223 support? I don't think Clojure or Scala currently implement that? Smalltalk isn't a scripting language - we actually compile everything to bytecodes before we execute. I think JSR 223 makes integration into existing applications easier. Also, many Frameworks make use of JSR 223 - e.g. Apache Pivot, Apache Camel or Apache ServiceMix. Wouldn't it be a great business-case to use Smalltalk on ServiceMix in an EAI or SOA Project? At least it would be very cool ;-) How do we define "production use"? The compiler runs today, simple scripts are running, James has some webapps running via Stout. I would say that if i can use redline just to "play around" with smalltalk on the JVM it wouldn't be ready for production. If i could call redline from an EJB and implement business logic -that would be production use. Or use it to write Unit-Tests for my Java-Application. regards, Stefan
|
|
|
Stevan, You can call java from Smalltalk and vice versa today, the process is currently a little manual (calling java from Smalltak) see latest Redline blog post on JVM byte codes To improve this we need a runtime adaptor for java classes. It's in the pipeline. I do need to document the process of getting the Smalltalk runtime going but it is just a sequence of method calls on Java classes Sent from my Commodore 64
|
|
|
In reply to this post by Steven! Ragnarök
On 29/04/12 8:10 PM, Steven! wrote:
> One last thing that would give us a really strong 1.0 would be a tool > like leiningen for Redline Smalltalk. For those unfamiliar, leiningen > is the Clojure build tool, one of the things that has made Clojure so > strong and given it really good penetration is the community's > emphasis that the clojure jar is just another library that you have on > your system and you can embed it in existing java apps and start out > writing strategic subsets in Clojure. Complementing this is leiningen > which is essentially the command line interface to Clojure. If anyone > is interested in writing this build tool with me (in Redline Smalltalk > of course) https://github.com/robertrolandorg/redline-maven-plugin I remembered it existed but for some reason thought it was part of the main project but not. Having tooling is certainly important, which is why I wrote the clojure-maven-plugin - leiningen is great, but sometimes you want to integrate with an existing project, and for that - maven support is key. For a 1.0 release I'd at least like to see the maven plugin become an official redline.st associated tool, released, and in maven central. |
|
|
I think Rob is onto maven support - Rob ?
Sent from my Commodore 64 On 02/05/2012, at 8:40 PM, Mark Derricutt <[hidden email]> wrote: > On 29/04/12 8:10 PM, Steven! wrote: >> One last thing that would give us a really strong 1.0 would be a tool like leiningen for Redline Smalltalk. For those unfamiliar, leiningen is the Clojure build tool, one of the things that has made Clojure so strong and given it really good penetration is the community's emphasis that the clojure jar is just another library that you have on your system and you can embed it in existing java apps and start out writing strategic subsets in Clojure. Complementing this is leiningen which is essentially the command line interface to Clojure. If anyone is interested in writing this build tool with me (in Redline Smalltalk of course) > Whats the status of the Maven plugin over at https://github.com/robertrolandorg/redline-maven-plugin > > I remembered it existed but for some reason thought it was part of the main project but not. > > Having tooling is certainly important, which is why I wrote the clojure-maven-plugin - leiningen is great, but sometimes you want to integrate with an existing project, and for that - maven support is key. > > For a 1.0 release I'd at least like to see the maven plugin become an official redline.st associated tool, released, and in maven central. |
|
|
In reply to this post by Mark Derricutt-2
Id also like to point out there is a ROADMAP document in the source that is a discussion point for release contents
Sent from my Commodore 64 On 02/05/2012, at 8:40 PM, Mark Derricutt <[hidden email]> wrote: > On 29/04/12 8:10 PM, Steven! wrote: >> One last thing that would give us a really strong 1.0 would be a tool like leiningen for Redline Smalltalk. For those unfamiliar, leiningen is the Clojure build tool, one of the things that has made Clojure so strong and given it really good penetration is the community's emphasis that the clojure jar is just another library that you have on your system and you can embed it in existing java apps and start out writing strategic subsets in Clojure. Complementing this is leiningen which is essentially the command line interface to Clojure. If anyone is interested in writing this build tool with me (in Redline Smalltalk of course) > Whats the status of the Maven plugin over at https://github.com/robertrolandorg/redline-maven-plugin > > I remembered it existed but for some reason thought it was part of the main project but not. > > Having tooling is certainly important, which is why I wrote the clojure-maven-plugin - leiningen is great, but sometimes you want to integrate with an existing project, and for that - maven support is key. > > For a 1.0 release I'd at least like to see the maven plugin become an official redline.st associated tool, released, and in maven central. |
|
|
In reply to this post by Mark Derricutt-2
The Maven plugin is an officially supported part of our tool chain. It exists within my github account simply because that's where I created it. :-) The main Redline account will keep up-to-date with it.
It has a chicken-and-egg problem - I can't release it to Maven Central until we have a Redline release to put in Maven Central. I'm also planning a Maven archetype. On May 2, 2012, at 3:42 AM, "Mark Derricutt" <[hidden email]> wrote: > On 29/04/12 8:10 PM, Steven! wrote: >> One last thing that would give us a really strong 1.0 would be a tool like leiningen for Redline Smalltalk. For those unfamiliar, leiningen is the Clojure build tool, one of the things that has made Clojure so strong and given it really good penetration is the community's emphasis that the clojure jar is just another library that you have on your system and you can embed it in existing java apps and start out writing strategic subsets in Clojure. Complementing this is leiningen which is essentially the command line interface to Clojure. If anyone is interested in writing this build tool with me (in Redline Smalltalk of course) > Whats the status of the Maven plugin over at https://github.com/robertrolandorg/redline-maven-plugin > > I remembered it existed but for some reason thought it was part of the main project but not. > > Having tooling is certainly important, which is why I wrote the clojure-maven-plugin - leiningen is great, but sometimes you want to integrate with an existing project, and for that - maven support is key. > > For a 1.0 release I'd at least like to see the maven plugin become an official redline.st associated tool, released, and in maven central. > |
«
Return to Redline Smalltalk
|
1 view|%1 views
| Free forum by Nabble | Edit this page |