January 2020 - Pharo-users - lists.pharo.org

Re: [Pharo-users] Why Smalltalk is so easy to evangelize

by jtuchel＠objektfabrik.de

Am 10.01.20 um 10:16 schrieb Marten Feldtmann: > That happened once in the history of Smalltalk and the big player was > IBM ... Well, twice actually ;-) Many people might not know that HP once was a Smalltalk vendor with their distributed Smalltalk (which was actually a white-label copy of VisualWorks iirc/iiuc)... Big corporations do not warrant the success of a technology. We'd still be using OS/2, BS2000, whatever...Â today if that was the case. The much bigger impact on Smalltalk's demise was that all vendors back in the nineties told their customers to switch somewehere else. One of them even never delivered what they had suggested as a replacement... I tried not to jump in here, but I am still a bit surprised how many people state that Smalltalk was so superior to (whatever hyped thing of the day). How many of us have built a nice iOS or Android App in Smalltalk? Do we even havy anything that might allow to do that in a power efficient way? How many of us have built a modern-looking SPA web application in Smalltalk? Smalltalk-only? These are areas where Smalltalk is quite weak and may not excel in any time soon. Let's be honest about it. None of the commercial vendors have announced anything in that direction, and to my knowledge, none of the open-source ones have anything more that the usual JS bridging. Dart/Flutter promises to finally be a platform on which you can write apps for both mobile platforms, the web and maybe even native with one single code base. Does that sound like a dream to developers who need to pick a tool? I guess so. Is there any point in telling people how great Smalltalk is when writing a mobile app or an SPA is hard to almost impossible? I know there are prototypes and even shipped apps on mobile platforms. Some are available and can be used, but are complex and mostly undocumented. One is private and doesn't even have a (public) price tag (yet?). Not really comparable to downloading Flutter and starting to write your first mobile app on a Saturday afternoon. Not sure the productivity advantage we keep praising (be it existent or not) will be noted much by somebody who needs to pick a tool now. So let's stop trying to convince people with things that mattered some 20 years ago. Even the function point thingie we keep carrying in front of our bellies (Capers-Jones was it?) is a lie when you want to build an application for today's markets. Smalltalk is great. If you don't need a mobile app to accompany your product. It's great if your GUI doesn't have to be sexy as hell or you are happy reinventing wheels. It is great if you only ship to PCs or on the web and don't need a lot of interaction in the browser. Anything else is hard in Smalltalk. If it's not, it is undocumented. I am not sure if energy spent on these "syntax fits on a postcard and, btw,Â we have the balloon" articles could be better spent doing something about the problems I mention here. I was in the same boat in the late 90ies and early 2000's with my blog and articles and stuff. All I found out was that nobody actually cares about these old hat stories. Heck, a lot of people these days don't even care about maintainability. You don't like it any more? No prob, we can redo it in this great new (JS) framework anyways. I am not saying anything is wrong about the maintainability of Smalltalk code or greatness of our IDE support. I like it and use it every day. I love programming in Smalltalk and hunting bugs in Smalltalk and fixing customers' problems in Smalltalk. I can fix a bug while the user tells me about theit problem and I am convinced not many other technologies can support me in this as good as Smalltalk does. But, unfortunately, I spend way too much time reinventing GUI glue code fo the web. And, unfortunately, I need to implement a lot of stuff in Javascript. Oh, and unfortunately, there is nothing as good as this if I wanted to ship a nice mobile app to augment our service. When it comes to these, I am not productive. I am not even close to any bleeding edge, and if I want to be, I need to do it on my own. Is it more productive to reinvent a Smalltalk version of VueJS just do be able to do the same as VueJS in JS than it would be to learn JS and VueJS and do some part of my work in those? I wanted to stay out of this thread, because it leads nowhere. But now that I've typed all this, I will push the send button and regret it in a few minutes... Joachim >and actually that really showed impact to the Smalltalk market. > Lots of consultings were running around, get pretty much money to teach > COBOL programmers how to use Smalltalk (or to be more precise: learn how > to click programs together). > > That hype perhaps lasted a few years ... and then IBM switched to Java > ... so they never can go back. > > > > > Marten > > Am 09.01.20 um 23:16 schrieb horrido: > >> It would be really nice to have some big tech company adopt Smalltalk, like >> Amazon, Apple, Facebook, Google, Microsoft, Netflix, Uber, etc. That would >> hit the ball right out of the park. Alas, I don't see that happening. I'm >> afraid JP Morgan, Siemens, and Thales aren't good enough. > -- ----------------------------------------------------------------------- Objektfabrik Joachim Tuchel mailto:jtuchel@objektfabrik.de Fliederweg 1 http://www.objektfabrik.de D-71640 Ludwigsburg http://joachimtuchel.wordpress.com Telefon: +49 7141 56 10 86 0 Fax: +49 7141 56 10 86 1

Jan. 10, 2020

Re: [Pharo-users] Why Smalltalk is so easy to evangelize

by Marten Feldtmann

That happened once in the history of Smalltalk and the big player was IBM ... and actually that really showed impact to the Smalltalk market. Lots of consultings were running around, get pretty much money to teach COBOL programmers how to use Smalltalk (or to be more precise: learn how to click programs together). That hype perhaps lasted a few years ... and then IBM switched to Java ... so they never can go back. Marten Am 09.01.20 um 23:16 schrieb horrido: > > It would be really nice to have some big tech company adopt Smalltalk, like > Amazon, Apple, Facebook, Google, Microsoft, Netflix, Uber, etc. That would > hit the ball right out of the park. Alas, I don't see that happening. I'm > afraid JP Morgan, Siemens, and Thales aren't good enough. -- Marten Feldtmann

Jan. 10, 2020

Re: [Pharo-users] Why Smalltalk is so easy to evangelize

by K K Subbu

On 10/01/20 12:25 AM, Kasper Ãsterbye wrote: > This rant states once again that in Smalltalk everything is an object. The word 'object' has been bandied about with multiple meanings, so it is understandable that you would challenge this claim. Smalltalk uses the term 'object' with a specific meaning - see chapter 30 of bluebook (Formal Specification of Object Memory)[1] for a concrete definition. The image is an object graph. The allocated memory in a heap consists entirely of a list objects which can be iterated through first/next messages. You need only two tools - Inspector and Explorer. Inspect to examine a single object and the explorer to trace interconnections in the object graph. Implementations like Squeak or Pharo may use strings instead of fully reifying programmable entities but that doesn't mean that they cannot. Also, some 'objects' may be managed entirely within the VM interpreter for pragmatic reasons. > To me, what really is nice about Smalltalk is NOT the language - it is > the image and live programming. And I can get around all the problems > with the language because of it. I miss: Bingo! Smalltalk is best understood as a virtual machine with live programming facility. Language is only a small part of it. [1] http://www.mirandabanda.org/bluebook/bluebook_chapter30.html Regards .. Subbu

Jan. 10, 2020

Re: [Pharo-users] Why Smalltalk is so easy to evangelize

by horrido

Oddly enough, I've had better results by appealing to history. I guess it has more to do with *how* I did it, my style and creativity. Things like Flutter and Elixir and Kotlin (for Android) are anomalies. Essentially, they benefitted from luck and word of mouth. You can't rely on that. While Smalltalk adoption has grown, if only slightly, it's still so far behind that much of the public continues to believe Smalltalk is dying. For me, that simply isn't good enough. It would be really nice to have some big tech company adopt Smalltalk, like Amazon, Apple, Facebook, Google, Microsoft, Netflix, Uber, etc. That would hit the ball right out of the park. Alas, I don't see that happening. I'm afraid JP Morgan, Siemens, and Thales aren't good enough. Esteban A. Maringolo wrote > Hi Richard, > > I don't find Smalltalk easy to evangelize, and in my experience the > appeal to history (a variation of the "argumentum ad antiquitatem" > fallacy) proved ineffective. > > People don't care about who invented MVC, bitblt or JIT, and so make > decisions looking into the future, they weight in the past of course, > but looking forward is what matters for any decision you take now. > > That's why things like Flutter or Elixir and other "new" technologies > get the attention they get these days, even when there are no "huge" > success cases. I can't judge whether these techs have value, are hyped > and/or there is a lot of FOMO in the decision making process. And no, > I don't believe it is because of Google shoving it through people > throats, it's people finding something valuable and trying to get an > professional advantage by learning/adopting it early. > > Smalltalk adoption in the last decade has grown by its own merits, > _despite_ of the efforts to promote it. > > I would bet that any appeal to emotion could be more effective, since > most developers get frustrated and any modern Smalltalk dialect can > ease that inherent frustration of software development, or even > better, turn it into an enjoyable experience (as it's been my case for > over a decade). > > Have some reasonable big tech/company saying they're going to use X, > and you'll have flocks of users trying X. > > Esteban A. Maringolo > > On Thu, Jan 9, 2020 at 5:03 PM horrido < > horrido.hobbies@ > > wrote: >> >> Absolutely correct. Each of those languages do have good reasons to >> choose >> them. I have never said otherwise. >> >> My point is that Smalltalk gives me many more reasons, many more ways to >> evangelize it. Smalltalk is very easy to evangelize. That's the premise >> of >> the entire article, and if it's wrong, then I should delete the entire >> article. >> >> Is it wrong? >> >> >> >> Esteban A. Maringolo wrote >> > On Thu, Jan 9, 2020 at 2:23 PM horrido < >> >> > horrido.hobbies@ >> >> > > wrote: >> > >> >> I happen to like Dart, Elixir, Golang, Julia, and Rust. But be honest: >> do >> >> these languages provide nearly as many reasons to choose them? >> >> I'm not being deprecatory. >> > >> > I don't know about Julia nor Elixir, but Dart has Flutter, Golang >> > drives a good chunk of the high-availability internet and Rust is >> > becoming the most secure programming language and several critical >> > applications are being rewritten in Rust. >> > >> > Their user base is huge (and so is their funding), but it's not only >> > about funding, the reasons to choose them are a lot, there is no >> > silver bullet. >> > >> > Regards, >> > >> > Esteban A. Maringolo >> >> >> >> >> >> -- >> Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html >> -- Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html

Jan. 9, 2020

Re: [Pharo-users] Why Smalltalk is so easy to evangelize

by Esteban Maringolo

Hi Richard, I don't find Smalltalk easy to evangelize, and in my experience the appeal to history (a variation of the "argumentum ad antiquitatem" fallacy) proved ineffective. People don't care about who invented MVC, bitblt or JIT, and so make decisions looking into the future, they weight in the past of course, but looking forward is what matters for any decision you take now. That's why things like Flutter or Elixir and other "new" technologies get the attention they get these days, even when there are no "huge" success cases. I can't judge whether these techs have value, are hyped and/or there is a lot of FOMO in the decision making process. And no, I don't believe it is because of Google shoving it through people throats, it's people finding something valuable and trying to get an professional advantage by learning/adopting it early. Smalltalk adoption in the last decade has grown by its own merits, _despite_ of the efforts to promote it. I would bet that any appeal to emotion could be more effective, since most developers get frustrated and any modern Smalltalk dialect can ease that inherent frustration of software development, or even better, turn it into an enjoyable experience (as it's been my case for over a decade). Have some reasonable big tech/company saying they're going to use X, and you'll have flocks of users trying X. Esteban A. Maringolo On Thu, Jan 9, 2020 at 5:03 PM horrido <horrido.hobbies(a)gmail.com> wrote: > > Absolutely correct. Each of those languages do have good reasons to choose > them. I have never said otherwise. > > My point is that Smalltalk gives me many more reasons, many more ways to > evangelize it. Smalltalk is very easy to evangelize. That's the premise of > the entire article, and if it's wrong, then I should delete the entire > article. > > Is it wrong? > > > > Esteban A. Maringolo wrote > > On Thu, Jan 9, 2020 at 2:23 PM horrido < > > > horrido.hobbies@ > > > > wrote: > > > >> I happen to like Dart, Elixir, Golang, Julia, and Rust. But be honest: do > >> these languages provide nearly as many reasons to choose them? > >> I'm not being deprecatory. > > > > I don't know about Julia nor Elixir, but Dart has Flutter, Golang > > drives a good chunk of the high-availability internet and Rust is > > becoming the most secure programming language and several critical > > applications are being rewritten in Rust. > > > > Their user base is huge (and so is their funding), but it's not only > > about funding, the reasons to choose them are a lot, there is no > > silver bullet. > > > > Regards, > > > > Esteban A. Maringolo > > > > > > -- > Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html >

Jan. 9, 2020

Re: [Pharo-users] Why Smalltalk is so easy to evangelize

by Sean P. DeNigris

Kasper Osterbye wrote > in Smalltalk everything is [not] an object > ... > * Message categories I agree with the thrust of your post and would like many of the items you suggest. That said, one semantic nitpick: "everything is an object" means as opposed to primitive types i.e. Date is an object that receives messages like any other, unlike in a language where it's a primitive type with hidden restrictions on how to interact with it. What you seem to be saying is that our objects are not always appropriate/optimal reflections of the domain, which is true but different. ----- Cheers, Sean -- Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html

Jan. 9, 2020

Re: [Pharo-users] Why Smalltalk is so easy to evangelize

by horrido

Which rant is that??? To me, what's really nice is the supremely simple language *and* the easily accessible programming environment *and* live coding *and* metaprogramming *and* the functional aspect (lambdas). It's not just one thing. It's the synergy that comes from the totality. However, there's no question that Smalltalk isn't perfect. There is certainly room for improvement, as you point out. But if there is a programming language that can come closer to perfection, I don't know what it is. And I've used a lot of languages. Such is the current state of affairs. Kasper Osterbye wrote > This rant states once again that in Smalltalk everything is an object. > Alas, it is not (but should). This is a shortlist of things which is > currently not objects in smalltalk: > > * Message categories > * Class categories (there is something called packages, which is rather > useful as they are actually objects, but they are not really done nicely - > manifest and RPackage???) > * The virtual machine (there is only one singleton thingie) > * The screen (there is only one why can one not instantiate more than one > world - I guess it is somewhat possible when we get gtk). > * Projects as (as in collection of objects). > > I happened to learn Simula before Smalltalk (I am Scandinavian after all). > When I program in Smalltalk I for sure miss nested classes and other > block-structured things. > > I happened to learn Beta (successor to Simula) before Smalltalk. I miss > being able to define virtual classes - but it is moot as there is no block > structure. > > To me, what really is nice about Smalltalk is NOT the language - it is the > image and live programming. And I can get around all the problems with the > language because of it. I miss: > > * Nested name spaces - when we finally get around to it, please do not do > just one level. > * Singular objects with behaviour I can write in a few lines (a single > object overriding one or a few methods) > * A simple switch/case statement > * â¦ > > But despite all this, I find programming in smalltalk much more fulfilling > than any other thing I ever touched. > > Best, > > Kasper -- Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html

Jan. 9, 2020

Re: [Pharo-users] Why Smalltalk is so easy to evangelize

by horrido

Absolutely correct. Each of those languages do have good reasons to choose them. I have never said otherwise. My point is that Smalltalk gives me many more reasons, many more ways to evangelize it. Smalltalk is very easy to evangelize. That's the premise of the entire article, and if it's wrong, then I should delete the entire article. Is it wrong? Esteban A. Maringolo wrote > On Thu, Jan 9, 2020 at 2:23 PM horrido < > horrido.hobbies@ > > wrote: > >> I happen to like Dart, Elixir, Golang, Julia, and Rust. But be honest: do >> these languages provide nearly as many reasons to choose them? >> I'm not being deprecatory. > > I don't know about Julia nor Elixir, but Dart has Flutter, Golang > drives a good chunk of the high-availability internet and Rust is > becoming the most secure programming language and several critical > applications are being rewritten in Rust. > > Their user base is huge (and so is their funding), but it's not only > about funding, the reasons to choose them are a lot, there is no > silver bullet. > > Regards, > > Esteban A. Maringolo -- Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html

Jan. 9, 2020

Re: [Pharo-users] Why Smalltalk is so easy to evangelize

by Kasper Ãsterbye

This rant states once again that in Smalltalk everything is an object. Alas, it is not (but should). This is a shortlist of things which is currently not objects in smalltalk: * Message categories * Class categories (there is something called packages, which is rather useful as they are actually objects, but they are not really done nicely - manifest and RPackage???) * The virtual machine (there is only one singleton thingie) * The screen (there is only one why can one not instantiate more than one world - I guess it is somewhat possible when we get gtk). * Projects as (as in collection of objects). I happened to learn Simula before Smalltalk (I am Scandinavian after all). When I program in Smalltalk I for sure miss nested classes and other block-structured things. I happened to learn Beta (successor to Simula) before Smalltalk. I miss being able to define virtual classes - but it is moot as there is no block structure. To me, what really is nice about Smalltalk is NOT the language - it is the image and live programming. And I can get around all the problems with the language because of it. I miss: * Nested name spaces - when we finally get around to it, please do not do just one level. * Singular objects with behaviour I can write in a few lines (a single object overriding one or a few methods) * A simple switch/case statement * â¦ But despite all this, I find programming in smalltalk much more fulfilling than any other thing I ever touched. Best, Kasper

Jan. 9, 2020

Re: [Pharo-users] Why Smalltalk is so easy to evangelize

by Esteban Maringolo

On Thu, Jan 9, 2020 at 2:23 PM horrido <horrido.hobbies(a)gmail.com> wrote: > I happen to like Dart, Elixir, Golang, Julia, and Rust. But be honest: do > these languages provide nearly as many reasons to choose them? > I'm not being deprecatory. I don't know about Julia nor Elixir, but Dart has Flutter, Golang drives a good chunk of the high-availability internet and Rust is becoming the most secure programming language and several critical applications are being rewritten in Rust. Their user base is huge (and so is their funding), but it's not only about funding, the reasons to choose them are a lot, there is no silver bullet. Regards, Esteban A. Maringolo

Jan. 9, 2020