Borland has anounced plans to buy TogetherSoft (http://www.togethersoft.com). Togethersoft produces an impressive integrated toolset for modeling tool for J2EE,.Net, XML, etc., round trip engineering, etc. Many of us prefer it to Rational's tools. Is this why MS is interested in Borland?
In addition to its stand-alone product, Togethersoft is scheduled to release a WebSphereStudio version in a few days. That version is built on eclipse (www.eclipse.org) -- an open source IDE used and developed by IBM.
All this would strengthen the MS position in traditional enterprises.
My father has found zoe from evectors pretty useful. Kind of google for email. He also has found mozilla to be a step up. In addition he has set up a wiki based on Squeak. And, obviously, uses Squeak. He is currenly looking at StarLogo 2.0.
My sister has found a wide assortment of VRML tools useful.
My mother has a hard time with most things these days, but she likes the idea of Squeak.
None of them have used apache, but they are almost ready.
While the implementation is Smalltalk specific, Niall Ross and
Andrew McQuiggin present an XP oriented approach to testing that may provide inspiration to slashdottes who can't make smalltalk. XML certainly is a possible approach for a non-Smalltalk implementation.
In addition, John Metsker's Building Parsers with Java includes a flexible approach to testing that goes well beyond traditional JUnit. In particular it addresses the issue of generating ad hoc tests that exercise unexpected input.
Big business under the microscope? Since the days of the Clinton administration, the SEC has not had enough money to read all the SEC filings. Anderson clearly sold accounting based upon a capacity to ignore bad news -- microscope or not.
The fact is that there have been many times in our history when business was under more real scrutiny. You know what? That is a major reason people all over the world started investing in US stocks.
Often CS programs are heavy on calculus and lite on discrete math so I'd include Concrete Mathematics (Graham, Knuth, Patashnik) plus Learning Discrete Mathematics with ISETL (Baxter, Dubinsky, Levin). The combination gives a reasonable combination of math and programming.
For learning programming as a discipline, I'd start with Structure and Interpretation of Computer Programs (Abelson's, Sussman, Sussman).
To gain a view of the implementation of a full computing system, Smalltalk-80: The Language and Its Implementation (Goldberg, Robson). Bytecodes to meta object protocol: this is a great worked example, placing emphasis on the critical distinction of language vs. implementation.
A great view of the modeling required for complex, yet flexible systems is Working With Objects (Reenskaug, the author of Model View Controller). Plus with this you can forget all the Swing books, unless there is separate value added. A key piece of Reenskaug is his notion of value chain. It turns out we are suppose to be adding value to someone else's work, not just cool hacks!
For Patterns, clearly Design Patterns is a required reference, but people need to be both smart and experienced to gain much from it. For learning Patterns I'd go to Design Patterns Java Workbook (Metsker), which also covers an an introduction to Swing and what passes for GUI programming. Saving time, we can order one more by Metsker: Building Parsers with Java, which not only presents a great framework for parsers (and a great tutorial by implication for framework design), but a strong introduction to unit testing, uml, and a greatly underappreciated framework for random testing. Since Metsker needs to parse something, logic programming, query processing, and imperative languages all get treatment that complements Structure and Interpretation of Computer Programs.
Is it really enough to cover everything with Scheme, ISETL, Smalltalk, and a little Java? Not quite. Partially because a good book on low level language will give something of an introduction to machine architecture. That sounds like C and I think The Practice of Programming (Kernighan, Pike) is far better than the standard white book. In addition, topics like incremental testing (unit tests without the xUnit brand), 'little languages', scripting languages, low level efficiency are introduced clearly. Continuing the path forward from C, Advanced C++ Programming Styles and Idioms (Coplien) is not just about C++, but about computer science.
Another reason, of course, for books in the C tradition is that they represent 'computing in the wild' - that is that's how a giant section of programmers works. (Of course, I'd say they are often doing it the hard way;) To finish with work in the C tradition, Effective Java is just fine. Far more than a Java book. It's good coverage of the problems of working with a language that reflects so many (I'm not saying poorly chosen) compromises.
Many workday issues are well covered in The Pragmatic Programmer (Thomas, Hunt).
Programming Perl (Wall, Schwartz)should be compared to Programming Ruby (Thomas, Hunt) and/or Programming Python (Lutz). The point is not to learn three scripting languages. The point is not to discover that my favorite is the best. The point is to gain an understanding of the differences and similarities among great programmers. More, to come to grips with the negative as well as positive consequences of almost any path.
Now there is this stuff called AI. This is not the place for an evaluation, but programmers need to know the work. Lets use google as our inspiration and go with Artificial Intelligence, A Modern Approach (Russell, Norvig). For an approach to AI from a different tradition, Genetic Programming (Koza). Watch out. This book is aggressive in its program. Something by Rodney Brooks belongs here, but I have not read his recent stuff and I can't find a copy of his earlier 'how to build a robot book). The point is that a work advocating a very bottom up, sensor/perception based model of computation/robotics/AI is needed to balance Russell and Norvig.
Data, Data, Data. Ullman's Database and Knowledge Base Systems covers a lot of bases. SQL gets placed in a sensible context. Swarm Intelligence (Kennedy, Eberhart) can round things out and set people up for Grid Computing Unleashed!. (author tbd).
Objects have been covered, so we need to take the next step and look at components. Here Component Software (Szyperski) is great. Some could say it's a bit dated, and it 5 years it will be, but right now it is solid and provides historical perspective on COM/CORBA, etc. Components are (largely) about deployment. The Java issues and opportunities are well covered in Component Development for the Java Platform (Halloway). The basic issues (dynamic class loading, security management, dynamic invocation, distribution, etc.) have strong are similar in C# (CLI, really), but I'd like a comparable book from that world.
Aspects and Generative programming are well introduced in Generative Programming: Methods, Tools, and Applications (Czarnecki, Eisenecker). For those who deal with (or should deal with) ontology, Appendix A is great. The discussion of Exemplar and Probabilistic ontology certainly raises some semantic web issues.
Writing code and changing it are pretty closely related. Fowlwer's Refactoring, is a great introduction to a disciplined approach to changing it. More: it can be read as a call for better tooling.
In order to have any hope of getting clear specifications and useful tests, programmers need to understand a bit of logic and probability. Language Proof and Logic (Barwise, Etchemendy) is great. It raises some deep issues concerning logic and visualization as well. I don't know what to recommend on the probability/statistics side. Certainly an introduction to Bayes should be included. Pearl's Causality or its predecessor exhaust my knowledge.
Graphics. Video. Sound. Music. 3d. Help!
Out of time, out of space. Distributed programming, operating systems, object databases, configuration management, source code control,and Linda (spaces) have not been mentioned.
I think this is one of the big advantages of XP's unit testing and pair programming. Sure, it can be over done, but the two in combination can really help people who 'get' the domain 'get' the spirit of the implementation.
In all honesty, introducing unit testing and pair programming is not easy for someone who has been primarly a coder on other projects. It is, however, possibly easier than a) doing everything yourself b) wasting time and money on consultants (like me!).
The next level to discuss is project management. Often good, realistic control over scope is part of regaining sanity in the project.
I'm not sure it is so simple. I belive many cs people at ga tech are pretty strong on collaboration -- or at least that is how I interpret their widespread use of
swiki and their very active participation in open source squeak .
I wonder if there is more to this story because some of the squeakers are hyperactive open source people.
Slashdot Mirror
Borland has anounced plans to buy TogetherSoft (http://www.togethersoft.com). Togethersoft produces an impressive integrated toolset for modeling tool for J2EE, .Net, XML, etc., round trip engineering, etc. Many of us prefer it to Rational's tools. Is this why MS is interested in Borland?
In addition to its stand-alone product, Togethersoft is scheduled to release a WebSphereStudio version in a few days. That version is built on eclipse (www.eclipse.org) -- an open source IDE used and developed by IBM.
All this would strengthen the MS position in traditional enterprises.
Nonesense. It would be worse if M$ bought the US Congress. No danger of that. Right?
My father has found zoe from evectors pretty useful. Kind of google for email. He also has found mozilla to be a step up. In addition he has set up a wiki based on Squeak. And, obviously, uses Squeak. He is currenly looking at StarLogo 2.0. My sister has found a wide assortment of VRML tools useful. My mother has a hard time with most things these days, but she likes the idea of Squeak. None of them have used apache, but they are almost ready.
While the implementation is Smalltalk specific, Niall Ross and Andrew McQuiggin present an XP oriented approach to testing that may provide inspiration to slashdottes who can't make smalltalk. XML certainly is a possible approach for a non-Smalltalk implementation. In addition, John Metsker's Building Parsers with Java includes a flexible approach to testing that goes well beyond traditional JUnit. In particular it addresses the issue of generating ad hoc tests that exercise unexpected input.
Big business under the microscope? Since the days of the Clinton administration, the SEC has not had enough money to read all the SEC filings. Anderson clearly sold accounting based upon a capacity to ignore bad news -- microscope or not. The fact is that there have been many times in our history when business was under more real scrutiny. You know what? That is a major reason people all over the world started investing in US stocks.
Often CS programs are heavy on calculus and lite on discrete math so I'd include Concrete Mathematics (Graham, Knuth, Patashnik) plus Learning Discrete Mathematics with ISETL (Baxter, Dubinsky, Levin). The combination gives a reasonable combination of math and programming. For learning programming as a discipline, I'd start with Structure and Interpretation of Computer Programs (Abelson's, Sussman, Sussman). To gain a view of the implementation of a full computing system, Smalltalk-80: The Language and Its Implementation (Goldberg, Robson). Bytecodes to meta object protocol: this is a great worked example, placing emphasis on the critical distinction of language vs. implementation. A great view of the modeling required for complex, yet flexible systems is Working With Objects (Reenskaug, the author of Model View Controller). Plus with this you can forget all the Swing books, unless there is separate value added. A key piece of Reenskaug is his notion of value chain. It turns out we are suppose to be adding value to someone else's work, not just cool hacks! For Patterns, clearly Design Patterns is a required reference, but people need to be both smart and experienced to gain much from it. For learning Patterns I'd go to Design Patterns Java Workbook (Metsker), which also covers an an introduction to Swing and what passes for GUI programming. Saving time, we can order one more by Metsker: Building Parsers with Java, which not only presents a great framework for parsers (and a great tutorial by implication for framework design), but a strong introduction to unit testing, uml, and a greatly underappreciated framework for random testing. Since Metsker needs to parse something, logic programming, query processing, and imperative languages all get treatment that complements Structure and Interpretation of Computer Programs. Is it really enough to cover everything with Scheme, ISETL, Smalltalk, and a little Java? Not quite. Partially because a good book on low level language will give something of an introduction to machine architecture. That sounds like C and I think The Practice of Programming (Kernighan, Pike) is far better than the standard white book. In addition, topics like incremental testing (unit tests without the xUnit brand), 'little languages', scripting languages, low level efficiency are introduced clearly. Continuing the path forward from C, Advanced C++ Programming Styles and Idioms (Coplien) is not just about C++, but about computer science. Another reason, of course, for books in the C tradition is that they represent 'computing in the wild' - that is that's how a giant section of programmers works. (Of course, I'd say they are often doing it the hard way;) To finish with work in the C tradition, Effective Java is just fine. Far more than a Java book. It's good coverage of the problems of working with a language that reflects so many (I'm not saying poorly chosen) compromises. Many workday issues are well covered in The Pragmatic Programmer (Thomas, Hunt). Programming Perl (Wall, Schwartz)should be compared to Programming Ruby (Thomas, Hunt) and/or Programming Python (Lutz). The point is not to learn three scripting languages. The point is not to discover that my favorite is the best. The point is to gain an understanding of the differences and similarities among great programmers. More, to come to grips with the negative as well as positive consequences of almost any path. Now there is this stuff called AI. This is not the place for an evaluation, but programmers need to know the work. Lets use google as our inspiration and go with Artificial Intelligence, A Modern Approach (Russell, Norvig). For an approach to AI from a different tradition, Genetic Programming (Koza). Watch out. This book is aggressive in its program. Something by Rodney Brooks belongs here, but I have not read his recent stuff and I can't find a copy of his earlier 'how to build a robot book). The point is that a work advocating a very bottom up, sensor/perception based model of computation/robotics/AI is needed to balance Russell and Norvig. Data, Data, Data. Ullman's Database and Knowledge Base Systems covers a lot of bases. SQL gets placed in a sensible context. Swarm Intelligence (Kennedy, Eberhart) can round things out and set people up for Grid Computing Unleashed!. (author tbd). Objects have been covered, so we need to take the next step and look at components. Here Component Software (Szyperski) is great. Some could say it's a bit dated, and it 5 years it will be, but right now it is solid and provides historical perspective on COM/CORBA, etc. Components are (largely) about deployment. The Java issues and opportunities are well covered in Component Development for the Java Platform (Halloway). The basic issues (dynamic class loading, security management, dynamic invocation, distribution, etc.) have strong are similar in C# (CLI, really), but I'd like a comparable book from that world. Aspects and Generative programming are well introduced in Generative Programming: Methods, Tools, and Applications (Czarnecki, Eisenecker). For those who deal with (or should deal with) ontology, Appendix A is great. The discussion of Exemplar and Probabilistic ontology certainly raises some semantic web issues. Writing code and changing it are pretty closely related. Fowlwer's Refactoring, is a great introduction to a disciplined approach to changing it. More: it can be read as a call for better tooling. In order to have any hope of getting clear specifications and useful tests, programmers need to understand a bit of logic and probability. Language Proof and Logic (Barwise, Etchemendy) is great. It raises some deep issues concerning logic and visualization as well. I don't know what to recommend on the probability/statistics side. Certainly an introduction to Bayes should be included. Pearl's Causality or its predecessor exhaust my knowledge. Graphics. Video. Sound. Music. 3d. Help! Out of time, out of space. Distributed programming, operating systems, object databases, configuration management, source code control,and Linda (spaces) have not been mentioned.
I think this is one of the big advantages of XP's unit testing and pair programming. Sure, it can be over done, but the two in combination can really help people who 'get' the domain 'get' the spirit of the implementation. In all honesty, introducing unit testing and pair programming is not easy for someone who has been primarly a coder on other projects. It is, however, possibly easier than a) doing everything yourself b) wasting time and money on consultants (like me!). The next level to discuss is project management. Often good, realistic control over scope is part of regaining sanity in the project.
I'm not sure it is so simple. I belive many cs people at ga tech are pretty strong on collaboration -- or at least that is how I interpret their widespread use of swiki and their very active participation in open source squeak . I wonder if there is more to this story because some of the squeakers are hyperactive open source people.