- Pharo-users - lists.pharo.org

Re: [Pharo-users] Smalltalk tutorial is broken

by James Foster

> On May 10, 2019, at 8:53 AM, Christopher Fuhrman <christopher.fuhrman(a)gmail.com> wrote: > > > On Fri, May 10, 2019, 15:32 Richard Kenneth Eng <horrido.hobbies(a)gmail.com <mailto:horrido.hobbies@gmail.com>> wrote: > So say many visitors to https://amber-lang.net/learn.html <https://amber-lang.net/learn.html>. > > The problem is this: visitors are failing to read and comprehend the last instruction on the first page of the tutorial. The one that reads: Select the text below and click on the 'DoIt' button. > ... > How to fix this? How to avoid this PR nightmare? > I can think of at least three easy solutions: > > Change the button from "DoIt" to "do the selected text" > > Change the first line to: > Profstef next "(you must select this entire line of text.)" > > But the best one might be: > Catch the parse error and if it's a null, remind the user she must first select the text. Some Smalltalk implementations auto-select the entire line if nothing is selected. I like that very much.

May 10, 2019

Re: [Pharo-users] Smalltalk tutorial is broken

by Ben Coman

A picture is worth a thousand words. What it needs is an animated GIF on loop demonstrating what to do. cheers -ben On Fri, 10 May 2019 at 21:32, Richard Kenneth Eng <horrido.hobbies(a)gmail.com> wrote: > > So say many visitors to https://amber-lang.net/learn.html. > > The problem is this: visitors are failing to read and comprehend the last instruction on the first page of the tutorial. The one that reads: Select the text below and click on the 'DoIt' button. > > They're not selecting the text. They do not understand what is meant by "select the text below." > > So they immediately encounter a parse error and conclude that the tutorial is broken. This is a horrible, horrible first impression for Smalltalk. > > It's not a rare occurrence, either. I've received this complaint numerous times on Quora and Medium. > > So the question is: How to fix this? How to avoid this PR nightmare?

May 10, 2019

Re: [Pharo-users] Smalltalk tutorial is broken

by Christopher Fuhrman

On Fri, May 10, 2019, 15:32 Richard Kenneth Eng <horrido.hobbies(a)gmail.com> wrote: > So say many visitors to https://amber-lang.net/learn.html. > > The problem is this: visitors are failing to read and comprehend the last > instruction on the first page of the tutorial. The one that reads: *Select > the text below and click on the 'DoIt' button*. > ... > How to fix this? How to avoid this PR nightmare? > I can think of at least three easy solutions: Change the button from "DoIt" to "do the selected text" Change the first line to: Profstef next "(you must select this entire line of text.)" But the best one might be: Catch the parse error and if it's a null, remind the user she must first select the text.

May 10, 2019

Re: [Pharo-users] Smalltalk tutorial is broken

by Jerry Kott

We have too many Bill Gatesâs and too few Steve Jobsâs. âuglification of the world - itâs a universal trendâ (Unbearable Lightness of Being, 1984). Smalltalk community suffers from too few people in general. For every 10000 âokâ worker you need one genius visionary to turn things upside down. Itâs always going to be a struggle. Feeling like being philosophical this morning... > On May 10, 2019, at 06:31, Richard Kenneth Eng <horrido.hobbies(a)gmail.com> wrote: > > So say many visitors to https://amber-lang.net/learn.html. > > The problem is this: visitors are failing to read and comprehend the last instruction on the first page of the tutorial. The one that reads: Select the text below and click on the 'DoIt' button. > > They're not selecting the text. They do not understand what is meant by "select the text below." > > So they immediately encounter a parse error and conclude that the tutorial is broken. This is a horrible, horrible first impression for Smalltalk. > > It's not a rare occurrence, either. I've received this complaint numerous times on Quora and Medium. > > So the question is: How to fix this? How to avoid this PR nightmare?

May 10, 2019

Fwd: introducing my GSoC project - Improving Code Completion

by Myroslava Romaniuk

[I guess the majority of people don't use the dev mailing list, so i'm resending this here.] Hi everyone, My name is Myroslava, I'm a 3rd year CS student in Ukraine, and my project Improving Code Completion has been selected for this year's GSoC. Some of you might know me from the 2 ESUGs I was at or maybe remember me previously doing GSoC in 2017, in any case I'm very happy to have been given this opportunity and I look forward to working with you! I started working on code completion last year during my internship at RMoD with Marcus and Stef, and I really want to make even more progress with code completion this summer and properly implement it inside Pharo as one of the options to use. You can read more about what has been done last summer and what is yet to be done this summer in my proposal (see attached). If you have any feedback or questions, don't hesitate to email me. Help and suggestions throughout my work on the project will be incredibly welcome too, as there's a lot to figure out how to do best. If you are interested in updates or want to keep in touch, I'm also on twitter <https://twitter.com/myroslavarm>, and I have a blog <https://medium.com/@myroslavarm> where I used to post about my progress in the last GSoC and about my internship at RMoD, and I will be writing there this summer, too. I'll also try to be more active in the mailing list and on discord :) Cheers, Myroslava

May 10, 2019

Smalltalk tutorial is broken

by Richard Kenneth Eng

So say many visitors to https://amber-lang.net/learn.html. The problem is this: visitors are failing to read and comprehend the last instruction on the first page of the tutorial. The one that reads: *Select the text below and click on the 'DoIt' button*. They're not selecting the text. They do not understand what is meant by "select the text below." So they immediately encounter a parse error and conclude that the tutorial is broken. This is a horrible, horrible first impression for Smalltalk. It's not a rare occurrence, either. I've received this complaint numerous times on Quora and Medium. So the question is: How to fix this? How to avoid this PR nightmare?

May 10, 2019

Re: [Pharo-users] GSoC 2019 Introduction

by Richard O'Keefe

Branching as such is not a sign of bad design in OOP. The actual dogma is that *testing the class of an argument instead of dispatching* is a sign of bad design. Just consider the case of binary search. That is not the kind of branching that lends itself to dynamic dispatch. Nor is the kind of branching you will have to write if you implement some sort of balanced tree. Or quad-trees or oct-trees or k-d trees or R-trees or ... I invite you to consider the following two methods from java.lang.String: byte[] getBytes(String charsetName); byte[] getBytes(CharSet charset); In Smalltalk, you have two options. The first is double dispatch. getBytes: encoding ^encoding getBytesFor: self String>>getBytesFor: CharSet>>getBytesFor: In this case, that requires you to modify two classes instead of one; in general it would require N+1 where N is the number of overloaded methods. It also leads to *coupling*, where too many classes depend on each other. Worst of all, in a project, you might not be *allowed* to modify CharSet. The other is to use an "if": getBytesFor: encoding |charset| charset := (encoding isKindOf: CharSet) ifTrue: [encoding] ifFalse: [CharSet named: encoding]. ... This approach - involves just one method, not N+1; - does not increase coupling; - does not require the ability to change system classes; - does not introduce additional non-private method names; - BUT does not cope well with additional ad hoc overloads. There is no silver bullet. Perhaps you may now look at Java with different eyes: ad hoc overloading is similar to using type-based branching. When it comes to data structures, I remember when we had a visitor to the CS department I was working in. His goal while he was with us was to carefully implement and benchmark about a dozen "advanced" priority queue data structures from the literature on the subject. I told him "make sure you include the classic array-based heap". Guess which algorithm won! That's right, the classic array-based heap. The point of the story is that a simple algorithm implemented well is likely to do surprisingly well. The rule of thumb I use is "keep the number of memory references down". As for which data structures you should look at, perhaps it is better to start by considering what *tasks* you want to support. If you want to support computational geometry, that will lead you to certain choices. If you want to support constraint satisfaction and/or combinatorial optimisation, that will lead you to others. If you want to make concurrent programming easier, that will lead you to concurrent stacks, queues, sets, bags, dictionaries &c. Don't think of "AVL tree", think of "efficient sorted sets, bags, and dictionaries" and work back from that. Here are two (related) data structures I've wanted. A triple store - is a set of (Att,Obj,Val) triples - supports *all* of the access patterns, where + means that value is known (given) and - means that value is unknown (iterated over) (+,+,+) (+,+,-) (+,-,+) (+,-,-) (-,+,+) (-,+,-) (-,-,+) (-,-,-) where each has good asymptotic efficiency - has low space overhead. This has obvious application to RDF and rule-based engines. A good graph data structure, as in nodes and edges. Well, graph AND digraph, obviously. You might want to look at Knuth's book "The Stanford GraphBase" for ideas. On Fri, 10 May 2019 at 22:00, Smiljana Knezev <smilja.knezev(a)gmail.com> wrote: > Hello everyone, > > I am a final year undergraduate student of Information Technologies in > Novi Sad, Serbia. I first heard about Pharo about a year ago when professor > Stephane Ducasse gave a guest lecture at our Faculty. I thought I was > familiar with Object Oriented Programming, having been learning Java, but > it was the first time I heard branching in OO is a sign of bad design. I > was drawn to Pharo because of it's simplicity and elegance. > > This Tuesday I found out I was accepted for GSoC program for New > Collections for Pharo project and I'm pretty excited about it. You can read > about proposal here: > https://summerofcode.withgoogle.com/projects/#4973054243897344 > It will focus on documenting, refactoring and testing existing collections > and migrating them to Containers library. Also some new collections like: > SimpleTrees, B-Trees, AVL Trees, QuadTrees, OcTrees, ... will be developed > Final part would be about research and development specialized/advanced > data structures like: Immutable Queues, Immutable Lists, Immutable > bidirectional maps,Threaded binary trees ... I would appreciate your > feedback about the proposal. Should I change something? Maybe add or focus > on some additional data structures? > > Other candidate from Serbia, Nina Medic and I discussed about writing a > blog together for our weekly posts. It can be found here: > https://smiljanaknezev.github.io/blog/?fbclid=IwAR2vC-N38wlPGY42ijokv67Anqd… > > It will be further customized and ready until Sunday. Also I have setup a > twitter account to promote our blog posts and keep in touch with community. > You can find me with @KnezevSmiljana . > > If you have any additional tips/tricks about making this GSoC project > successful, I would like to hear about them. > > Looking forward to hearing from you. > > Best regards, > Smiljana > > > > >

May 10, 2019

GSoC 2019 introducing my project: Next Generation of Unit Testing

by Dayne Lorena Guerra Calle

Hello Everyone, My name is Dayne, I'm a final-year computer science student in Bolivia and I have been selected with my project as one of the students for this GSOC. First I will work in DrTests: DrTests project aims to provide a plugin-based UI to deal with tests in Pharo. It will provide the same features as the actual SUnit UI (i.e., running tests, profiling tests and computing code-coverage) but will allow to plug additional analysis on unit tests. If you want to try the first version of DrTests in Pharo 7.0: https://github.com/juliendelplanque/DrTests Feedback and suggestions are always welcome. Now I am migrating the UI from spec to spec2 for DrTest to be available in Pharo 8.0 Then the plan is enhancing the SUnit implementation by introducing a common Pharo's Sunit layer instead of different layer doing the same (for example, the SUnit UI, Jenkins tools and the system browser define 3 different ways to collect tests defined in a package). And If you want to know more details bout this project I will attach my proposal. I will write posts about the work done in my blog <http://medium.com/@daynestorm> and you can contact me If you have new ideas for plugins or suggestions :) Cheers, Dayne

May 10, 2019

GSoC 2019 Introduction

by Smiljana Knezev

Hello everyone, I am a final year undergraduate student of Information Technologies in Novi Sad, Serbia. I first heard about Pharo about a year ago when professor Stephane Ducasse gave a guest lecture at our Faculty. I thought I was familiar with Object Oriented Programming, having been learning Java, but it was the first time I heard branching in OO is a sign of bad design. I was drawn to Pharo because of it's simplicity and elegance. This Tuesday I found out I was accepted for GSoC program for New Collections for Pharo project and I'm pretty excited about it. You can read about proposal here: https://summerofcode.withgoogle.com/projects/#4973054243897344 It will focus on documenting, refactoring and testing existing collections and migrating them to Containers library. Also some new collections like: SimpleTrees, B-Trees, AVL Trees, QuadTrees, OcTrees, ... will be developed Final part would be about research and development specialized/advanced data structures like: Immutable Queues, Immutable Lists, Immutable bidirectional maps,Threaded binary trees ... I would appreciate your feedback about the proposal. Should I change something? Maybe add or focus on some additional data structures? Other candidate from Serbia, Nina Medic and I discussed about writing a blog together for our weekly posts. It can be found here: https://smiljanaknezev.github.io/blog/?fbclid=IwAR2vC-N38wlPGY42ijokv67Anqd… It will be further customized and ready until Sunday. Also I have setup a twitter account to promote our blog posts and keep in touch with community. You can find me with @KnezevSmiljana . If you have any additional tips/tricks about making this GSoC project successful, I would like to hear about them. Looking forward to hearing from you. Best regards, Smiljana

May 10, 2019

GSoC 2019 Introduction myself

by EVELYN CUSI LOPEZ

Hello, everybody. My name is Evelyn, I'm 24 years old. I'm a last year student at the Mayor de San Simon University in Bolivia, I'm currently doing my degree project to get my professional degree. My proposal for "Better and more refactorings for Pharo" was selected in this year's GsOC :D. Just few months ago I got to know the world of Smalltalk and I really liked it, I made a page for ESUG in Seaside and I made some extensions to use Magritte with Material Design Lite, for now I'm doing a tutorial of Parasol still missing some details but soon will be ready. As for my proposal is to improve the current refactoring tools that Pharo currently has, I plan to add some shortcuts to the most important refactoring options, add some previews before refactoring similar to IntelliJ IDEA, add three new refactoring options (Inline method, variable inline, and find and replace) and finally improve the option to extract method by adding a duplicate code search of the method that will be extracted. Now I am learning everything necessary for the realization of this project, so any suggestion to improve the proposal or suggest the best way to start is welcome, and do not hesitate to ask any questions you have about this project :D. I'm very excited to start working with you. Sincerely, Evelyn. *P.d.* You can access my proposal in the attached file: https://storage.googleapis.com/summerofcode-prod.appspot.com/gsoc/core_proj…

May 10, 2019