Thread overview | ||||||||
---|---|---|---|---|---|---|---|---|
|
December 25, 2007 Merry Christmas, and an STLSoft New Year! | ||||
---|---|---|---|---|
| ||||
First, I'd like to wish everyone a healthy and happy holiday. And I'd also like to let you know what you're in for from me with STLSoft (and related projects) in 2008. So, in no particular order: * STLSoft: 1. Website. This is finally going to get the professional treatment. It will have a redesign, be moved to its own server (so stlsoft.org is no longer synesis.com.au/software/stlsoft), have a blog, have a place for publishing small (or even large) articles, have tutorials, and a whole lot more. This should happen in Jan/Feb. Later in the year I may also provide commercial support facilities. 2. Documentation. The first phase will be to get the STLSoft distribution scripts updated and to start re-publishing the Doxygen generated docs. This should occur in Jan. In a couple more months' time, once the site re-org is done, I will be providing better and more extensive documentation, in concert with some other "senior" STLSoft fellows. More on this at another time ... 3. 1.10 will be released soon. It will contain several new components - including an uber-efficient properties-file class - and maybe even a couple of new projects (of which more later). One important thing is for me to address all the remaining items on the NG: there are currently 62 marked unread (i.e. left to process), some of which go back to 2005! Finally, there'll be a first tranch of some much needed rationalisation of internal/support files/classes/namespaces, to simplify and reduce compilation times. 4. Later in the year I plan, if I get time, to release STLSoft+, which is a commercial enhancement to STLSoft, supplementing various STLSoft sub-projects with substantial services. 5. Support for other new libraries, including xContract (see below) * Synesis: 1. The website's going to get a radical reworking (and reduction), to reflect the commercial emphases of the last few years: development process consulting, open-source customisation, network-related custom product development. 2. All the to-be-released soon (some as back as 2004!) will be removed. 3. The system tools will be updated (for Linux & Windows, 32 & 64-bit) 4. Some articles and/or blogs * "Breaking Up The Monolith: Advanced C++ Design without Compromise" (http://breakingupthemonolith.com/) This is my next book, and is about how to get everything you want from C++ without compromise, particularly between efficiency, expressiveness and robustness. It discusses the technologies (in particular the Shims concept and the Type Tunnel pattern) behind Pantheios and FastFormat, as well as other techniques for breaking up monolithic software, maximising design values, and so on. I've been writing it on and off for the last 15 months, but am about to put in a solid 10 days on it, which should get me to the 50% completion mark. The rest will be done over weekends over the following 3 months. For anyone that's followed Imperfect C++ & Extended STL (vol 1) and/or the progression of the STLSoft, Pantheios, recls, FastFormat, flecxx, VOLE etc. libraries, this is the book that explains it all, including the Shims concept, the Handle::Ref+Wrapper and Type Tunnel patterns, the principles of Intersecting Conformance, Irrecoverability and Removeability, and more. Should be out mid/late 2008. * Pantheios As you may know, this is, IMO, the ultimate logging API library, affording the user complete robustness (incl. 100% type-safety) and unmatchable performance characteristics. It is designed to work above any other logging libraries, so you can have maximal performance combined with the configurability of, say, log4cxx. Pantheios has been in beta for over a year, but is now very close to a final 1.0 release. The remaining tasks are: 1. put in xTests (see below) unit and component tests into the build 2. implement buffering and file-rolling for be.file 3. website rework 4. project files for VC++ 2003/5/8 & Borland Turbo C++ 5. update documentation 6. compatibility with VC++ "safe" compilation I've also been involved with some commercial customisations of Pantheios for companies here in Aus and in the US, who believe that Pantheios offers them a significant competitive advantage. Who am I to disagree? <g> I can't comment on what I've been doing, or for whom I've been doing it, but I am, of course, always available for such work for any clients that believe their logging requirements are such that they have to eek out the maximum value of every last processor cycle in the software. * FastFormat FastFormat uses a similar technology to Pantheios to afford total type-safety and unbeatable performance when formatting text statements. It fully supports I18N/L10N by the use of numbered arguments; arguments used more than once are converted only once. It supports arbitrary destinations of the formatted result by the mechanism of "sinks". There are several stock sink types, including strings, string arrays, files, speech (currently Windows only). FF is almost ready to go - I've built the makefile template and proved it on Linux/Mac/Windows with several compilers - and should be out next week. Am just working through some other stuff first, and then plan to write the FF chapter for "Monolith" at the time that I release it (and submit article(s) on it to mags). Although my publicity machine has never been the best, I believe that there should be no reason why FF should not become the de facto standard for C++ formatting/output, as it addresses all the problems of C's streams, C++'s IOStreams, and the other open-source "modern" formatting libraries. Time will tell ... ;-) * flecxx One of the more simplistic messages in "Monolith" is that good libraries should afford one the ability to communicate with them in the types with which you're writing your application code, rather than the lower-level-of-abstraction types in which their interfaces are expressed. flecxx, to be released in Jan/Feb, is a 100% header-only adaptation layer library that wraps standard and popular 3rd-party libraries and removes the need to sully your beautiful code with all those .get(), .c_str(), .GetSafeHWnd(), etc. calls. * xTests After releasing all these libraries, it's about time I put out the testing framework I've been using these long years. This is particularly so because the next release of Pantheios, and every release of FastFormat and flecxx, will come bundled with the testing framework. Open-sourced and rebadged as xTests, it'll be available for anyone to use with any projects sometime soon. xTests is deliberately not all that sophisticated, so it's not going to be a replacement for any xUnit-like library you're using. But it's simple, and has minimal coupling, and does the job. * xContract Testing's important, but it's only half the software quality assurance picture.Contracts are the other. I'm going to put my money where my mouth is this year, and reify all my wild theories about Irrecoverability in the form of the xContracts library. We'll start with C/C++, then Ruby, and then see where life takes us after that. That's a brief summary of my non-commercial effort over the coming year. All those people who've been understandably frustrated about the lack of doc/support with STLSoft should, by the end of '08, have all their issues dequalmed. A happy new year to all! :-) Matt |
December 27, 2007 Re: Merry Christmas, and an STLSoft New Year! | ||||
---|---|---|---|---|
| ||||
Posted in reply to Matthew Wilson | Go go Matt!!
You inspire to aspire!
Keep up the great work,
Adi
Matthew Wilson wrote:
> First, I'd like to wish everyone a healthy and happy holiday.
>
>
> And I'd also like to let you know what you're in for from me with STLSoft
> (and related projects) in 2008. So, in no particular order:
>
> * STLSoft:
> 1. Website. This is finally going to get the professional treatment. It
> will have a redesign, be moved to its own server (so stlsoft.org is no
> longer synesis.com.au/software/stlsoft), have a blog, have a place for
> publishing small (or even large) articles, have tutorials, and a whole lot
> more. This should happen in Jan/Feb. Later in the year I may also provide
> commercial support facilities.
> 2. Documentation. The first phase will be to get the STLSoft
> distribution scripts updated and to start re-publishing the Doxygen
> generated docs. This should occur in Jan. In a couple more months' time,
> once the site re-org is done, I will be providing better and more extensive
> documentation, in concert with some other "senior" STLSoft fellows. More on
> this at another time ...
> 3. 1.10 will be released soon. It will contain several new components -
> including an uber-efficient properties-file class - and maybe even a couple
> of new projects (of which more later). One important thing is for me to
> address all the remaining items on the NG: there are currently 62 marked
> unread (i.e. left to process), some of which go back to 2005! Finally,
> there'll be a first tranch of some much needed rationalisation of
> internal/support files/classes/namespaces, to simplify and reduce
> compilation times.
> 4. Later in the year I plan, if I get time, to release STLSoft+, which
> is a commercial enhancement to STLSoft, supplementing various STLSoft
> sub-projects with substantial services.
> 5. Support for other new libraries, including xContract (see below)
>
>
> * Synesis:
> 1. The website's going to get a radical reworking (and reduction), to
> reflect the commercial emphases of the last few years: development process
> consulting, open-source customisation, network-related custom product
> development.
> 2. All the to-be-released soon (some as back as 2004!) will be removed.
> 3. The system tools will be updated (for Linux & Windows, 32 & 64-bit)
> 4. Some articles and/or blogs
>
>
> * "Breaking Up The Monolith: Advanced C++ Design without Compromise"
> (http://breakingupthemonolith.com/)
> This is my next book, and is about how to get everything you want from
> C++ without compromise, particularly between efficiency, expressiveness and
> robustness. It discusses the technologies (in particular the Shims concept
> and the Type Tunnel pattern) behind Pantheios and FastFormat, as well as
> other techniques for breaking up monolithic software, maximising design
> values, and so on. I've been writing it on and off for the last 15 months,
> but am about to put in a solid 10 days on it, which should get me to the 50%
> completion mark. The rest will be done over weekends over the following 3
> months. For anyone that's followed Imperfect C++ & Extended STL (vol 1)
> and/or the progression of the STLSoft, Pantheios, recls, FastFormat, flecxx,
> VOLE etc. libraries, this is the book that explains it all, including the
> Shims concept, the Handle::Ref+Wrapper and Type Tunnel patterns, the
> principles of Intersecting Conformance, Irrecoverability and Removeability,
> and more. Should be out mid/late 2008.
>
>
> * Pantheios
> As you may know, this is, IMO, the ultimate logging API library,
> affording the user complete robustness (incl. 100% type-safety) and
> unmatchable performance characteristics. It is designed to work above any
> other logging libraries, so you can have maximal performance combined with
> the configurability of, say, log4cxx.
> Pantheios has been in beta for over a year, but is now very close to a
> final 1.0 release. The remaining tasks are:
> 1. put in xTests (see below) unit and component tests into the build
> 2. implement buffering and file-rolling for be.file
> 3. website rework
> 4. project files for VC++ 2003/5/8 & Borland Turbo C++
> 5. update documentation
> 6. compatibility with VC++ "safe" compilation
> I've also been involved with some commercial customisations of Pantheios
> for companies here in Aus and in the US, who believe that Pantheios offers
> them a significant competitive advantage. Who am I to disagree? <g> I can't
> comment on what I've been doing, or for whom I've been doing it, but I am,
> of course, always available for such work for any clients that believe their
> logging requirements are such that they have to eek out the maximum value of
> every last processor cycle in the software.
>
>
> * FastFormat
> FastFormat uses a similar technology to Pantheios to afford total
> type-safety and unbeatable performance when formatting text statements. It
> fully supports I18N/L10N by the use of numbered arguments; arguments used
> more than once are converted only once. It supports arbitrary destinations
> of the formatted result by the mechanism of "sinks". There are several stock
> sink types, including strings, string arrays, files, speech (currently
> Windows only).
> FF is almost ready to go - I've built the makefile template and proved
> it on Linux/Mac/Windows with several compilers - and should be out next
> week. Am just working through some other stuff first, and then plan to write
> the FF chapter for "Monolith" at the time that I release it (and submit
> article(s) on it to mags).
> Although my publicity machine has never been the best, I believe that
> there should be no reason why FF should not become the de facto standard for
> C++ formatting/output, as it addresses all the problems of C's streams,
> C++'s IOStreams, and the other open-source "modern" formatting libraries.
> Time will tell ... ;-)
>
>
> * flecxx
>
> One of the more simplistic messages in "Monolith" is that good libraries
> should afford one the ability to communicate with them in the types with
> which you're writing your application code, rather than the
> lower-level-of-abstraction types in which their interfaces are expressed.
> flecxx, to be released in Jan/Feb, is a 100% header-only adaptation layer
> library that wraps standard and popular 3rd-party libraries and removes the
> need to sully your beautiful code with all those .get(), .c_str(),
> .GetSafeHWnd(), etc. calls.
>
>
> * xTests
> After releasing all these libraries, it's about time I put out the
> testing framework I've been using these long years. This is particularly so
> because the next release of Pantheios, and every release of FastFormat and
> flecxx, will come bundled with the testing framework. Open-sourced and
> rebadged as xTests, it'll be available for anyone to use with any projects
> sometime soon.
> xTests is deliberately not all that sophisticated, so it's not going to
> be a replacement for any xUnit-like library you're using. But it's simple,
> and has minimal coupling, and does the job.
>
>
> * xContract
> Testing's important, but it's only half the software quality assurance
> picture.Contracts are the other. I'm going to put my money where my mouth is
> this year, and reify all my wild theories about Irrecoverability in the form
> of the xContracts library. We'll start with C/C++, then Ruby, and then see
> where life takes us after that.
>
>
>
> That's a brief summary of my non-commercial effort over the coming year. All
> those people who've been understandably frustrated about the lack of
> doc/support with STLSoft should, by the end of '08, have all their issues
> dequalmed.
>
> A happy new year to all!
>
> :-)
>
> Matt
>
>
>
|
December 29, 2007 Re: Merry Christmas, and an STLSoft New Year! | ||||
---|---|---|---|---|
| ||||
Posted in reply to Adi Shavit | Thanks mate, you old charmer ... ;-) "Adi Shavit" <adish@gentech.co.il> wrote in message news:4774095C.1020202@gentech.co.il... > Go go Matt!! > You inspire to aspire! > Keep up the great work, > Adi > > Matthew Wilson wrote: > > First, I'd like to wish everyone a healthy and happy holiday. > > > > > > And I'd also like to let you know what you're in for from me with STLSoft > > (and related projects) in 2008. So, in no particular order: > > > > * STLSoft: > > 1. Website. This is finally going to get the professional treatment. It > > will have a redesign, be moved to its own server (so stlsoft.org is no longer synesis.com.au/software/stlsoft), have a blog, have a place for publishing small (or even large) articles, have tutorials, and a whole lot > > more. This should happen in Jan/Feb. Later in the year I may also provide > > commercial support facilities. > > 2. Documentation. The first phase will be to get the STLSoft > > distribution scripts updated and to start re-publishing the Doxygen > > generated docs. This should occur in Jan. In a couple more months' time, > > once the site re-org is done, I will be providing better and more extensive > > documentation, in concert with some other "senior" STLSoft fellows. More on > > this at another time ... > > 3. 1.10 will be released soon. It will contain several new components - > > including an uber-efficient properties-file class - and maybe even a couple > > of new projects (of which more later). One important thing is for me to > > address all the remaining items on the NG: there are currently 62 marked > > unread (i.e. left to process), some of which go back to 2005! Finally, > > there'll be a first tranch of some much needed rationalisation of > > internal/support files/classes/namespaces, to simplify and reduce > > compilation times. > > 4. Later in the year I plan, if I get time, to release STLSoft+, which > > is a commercial enhancement to STLSoft, supplementing various STLSoft > > sub-projects with substantial services. > > 5. Support for other new libraries, including xContract (see below) > > > > > > * Synesis: > > 1. The website's going to get a radical reworking (and reduction), to > > reflect the commercial emphases of the last few years: development process > > consulting, open-source customisation, network-related custom product > > development. > > 2. All the to-be-released soon (some as back as 2004!) will be removed. > > 3. The system tools will be updated (for Linux & Windows, 32 & 64-bit) > > 4. Some articles and/or blogs > > > > > > * "Breaking Up The Monolith: Advanced C++ Design without Compromise" > > (http://breakingupthemonolith.com/) > > This is my next book, and is about how to get everything you want from > > C++ without compromise, particularly between efficiency, expressiveness and > > robustness. It discusses the technologies (in particular the Shims concept > > and the Type Tunnel pattern) behind Pantheios and FastFormat, as well as other techniques for breaking up monolithic software, maximising design values, and so on. I've been writing it on and off for the last 15 months, > > but am about to put in a solid 10 days on it, which should get me to the 50% > > completion mark. The rest will be done over weekends over the following 3 > > months. For anyone that's followed Imperfect C++ & Extended STL (vol 1) and/or the progression of the STLSoft, Pantheios, recls, FastFormat, flecxx, > > VOLE etc. libraries, this is the book that explains it all, including the > > Shims concept, the Handle::Ref+Wrapper and Type Tunnel patterns, the principles of Intersecting Conformance, Irrecoverability and Removeability, > > and more. Should be out mid/late 2008. > > > > > > * Pantheios > > As you may know, this is, IMO, the ultimate logging API library, > > affording the user complete robustness (incl. 100% type-safety) and > > unmatchable performance characteristics. It is designed to work above any > > other logging libraries, so you can have maximal performance combined with > > the configurability of, say, log4cxx. > > Pantheios has been in beta for over a year, but is now very close to a > > final 1.0 release. The remaining tasks are: > > 1. put in xTests (see below) unit and component tests into the build > > 2. implement buffering and file-rolling for be.file > > 3. website rework > > 4. project files for VC++ 2003/5/8 & Borland Turbo C++ > > 5. update documentation > > 6. compatibility with VC++ "safe" compilation > > I've also been involved with some commercial customisations of Pantheios > > for companies here in Aus and in the US, who believe that Pantheios offers > > them a significant competitive advantage. Who am I to disagree? <g> I can't > > comment on what I've been doing, or for whom I've been doing it, but I am, > > of course, always available for such work for any clients that believe their > > logging requirements are such that they have to eek out the maximum value of > > every last processor cycle in the software. > > > > > > * FastFormat > > FastFormat uses a similar technology to Pantheios to afford total > > type-safety and unbeatable performance when formatting text statements. It > > fully supports I18N/L10N by the use of numbered arguments; arguments used > > more than once are converted only once. It supports arbitrary destinations > > of the formatted result by the mechanism of "sinks". There are several stock > > sink types, including strings, string arrays, files, speech (currently > > Windows only). > > FF is almost ready to go - I've built the makefile template and proved > > it on Linux/Mac/Windows with several compilers - and should be out next week. Am just working through some other stuff first, and then plan to write > > the FF chapter for "Monolith" at the time that I release it (and submit > > article(s) on it to mags). > > Although my publicity machine has never been the best, I believe that > > there should be no reason why FF should not become the de facto standard for > > C++ formatting/output, as it addresses all the problems of C's streams, C++'s IOStreams, and the other open-source "modern" formatting libraries. > > Time will tell ... ;-) > > > > > > * flecxx > > > > One of the more simplistic messages in "Monolith" is that good libraries > > should afford one the ability to communicate with them in the types with which you're writing your application code, rather than the lower-level-of-abstraction types in which their interfaces are expressed. > > flecxx, to be released in Jan/Feb, is a 100% header-only adaptation layer > > library that wraps standard and popular 3rd-party libraries and removes the > > need to sully your beautiful code with all those .get(), .c_str(), > > .GetSafeHWnd(), etc. calls. > > > > > > * xTests > > After releasing all these libraries, it's about time I put out the > > testing framework I've been using these long years. This is particularly so > > because the next release of Pantheios, and every release of FastFormat and > > flecxx, will come bundled with the testing framework. Open-sourced and rebadged as xTests, it'll be available for anyone to use with any projects > > sometime soon. > > xTests is deliberately not all that sophisticated, so it's not going to > > be a replacement for any xUnit-like library you're using. But it's simple, > > and has minimal coupling, and does the job. > > > > > > * xContract > > Testing's important, but it's only half the software quality assurance > > picture.Contracts are the other. I'm going to put my money where my mouth is > > this year, and reify all my wild theories about Irrecoverability in the form > > of the xContracts library. We'll start with C/C++, then Ruby, and then see > > where life takes us after that. > > > > > > > > That's a brief summary of my non-commercial effort over the coming year. All > > those people who've been understandably frustrated about the lack of doc/support with STLSoft should, by the end of '08, have all their issues > > dequalmed. > > > > A happy new year to all! > > > > :-) > > > > Matt > > > > > > |
December 30, 2007 Re: Merry Christmas, and an STLSoft New Year! | ||||
---|---|---|---|---|
| ||||
Posted in reply to Matthew Wilson | Hi Mathew, Beforehand happy holidays! Without any doubt I have to agree with Mr. Adi Shavit, I will be honest that the thing that amazes me the most is the level of "production" you present and sustain (synesis + book + stlsoft + personal life + ... ) my daily job alone "sucks" the life out of me. Anyway, could you give some detail regarding STLSoft+, since you detailed all the other stuff. Happy New Year Cláudio Albuquerque "Matthew Wilson" <matthew@hat.stlsoft.dot.org> wrote in message news:fks1rp$1uu7$1@digitalmars.com... > First, I'd like to wish everyone a healthy and happy holiday. > > > And I'd also like to let you know what you're in for from me with STLSoft (and related projects) in 2008. So, in no particular order: > > * STLSoft: > 1. Website. This is finally going to get the professional treatment. It > will have a redesign, be moved to its own server (so stlsoft.org is no > longer synesis.com.au/software/stlsoft), have a blog, have a place for > publishing small (or even large) articles, have tutorials, and a whole lot > more. This should happen in Jan/Feb. Later in the year I may also provide > commercial support facilities. > 2. Documentation. The first phase will be to get the STLSoft > distribution scripts updated and to start re-publishing the Doxygen > generated docs. This should occur in Jan. In a couple more months' time, > once the site re-org is done, I will be providing better and more > extensive > documentation, in concert with some other "senior" STLSoft fellows. More > on > this at another time ... > 3. 1.10 will be released soon. It will contain several new components - > including an uber-efficient properties-file class - and maybe even a > couple > of new projects (of which more later). One important thing is for me to > address all the remaining items on the NG: there are currently 62 marked > unread (i.e. left to process), some of which go back to 2005! Finally, > there'll be a first tranch of some much needed rationalisation of > internal/support files/classes/namespaces, to simplify and reduce > compilation times. > 4. Later in the year I plan, if I get time, to release STLSoft+, which > is a commercial enhancement to STLSoft, supplementing various STLSoft > sub-projects with substantial services. > 5. Support for other new libraries, including xContract (see below) > > > * Synesis: > 1. The website's going to get a radical reworking (and reduction), to > reflect the commercial emphases of the last few years: development process > consulting, open-source customisation, network-related custom product > development. > 2. All the to-be-released soon (some as back as 2004!) will be removed. > 3. The system tools will be updated (for Linux & Windows, 32 & 64-bit) > 4. Some articles and/or blogs > > > * "Breaking Up The Monolith: Advanced C++ Design without Compromise" > (http://breakingupthemonolith.com/) > This is my next book, and is about how to get everything you want from > C++ without compromise, particularly between efficiency, expressiveness > and > robustness. It discusses the technologies (in particular the Shims concept > and the Type Tunnel pattern) behind Pantheios and FastFormat, as well as > other techniques for breaking up monolithic software, maximising design > values, and so on. I've been writing it on and off for the last 15 months, > but am about to put in a solid 10 days on it, which should get me to the > 50% > completion mark. The rest will be done over weekends over the following 3 > months. For anyone that's followed Imperfect C++ & Extended STL (vol 1) > and/or the progression of the STLSoft, Pantheios, recls, FastFormat, > flecxx, > VOLE etc. libraries, this is the book that explains it all, including the > Shims concept, the Handle::Ref+Wrapper and Type Tunnel patterns, the > principles of Intersecting Conformance, Irrecoverability and > Removeability, > and more. Should be out mid/late 2008. > > > * Pantheios > As you may know, this is, IMO, the ultimate logging API library, > affording the user complete robustness (incl. 100% type-safety) and > unmatchable performance characteristics. It is designed to work above any > other logging libraries, so you can have maximal performance combined with > the configurability of, say, log4cxx. > Pantheios has been in beta for over a year, but is now very close to a > final 1.0 release. The remaining tasks are: > 1. put in xTests (see below) unit and component tests into the > build > 2. implement buffering and file-rolling for be.file > 3. website rework > 4. project files for VC++ 2003/5/8 & Borland Turbo C++ > 5. update documentation > 6. compatibility with VC++ "safe" compilation > I've also been involved with some commercial customisations of > Pantheios > for companies here in Aus and in the US, who believe that Pantheios offers > them a significant competitive advantage. Who am I to disagree? <g> I > can't > comment on what I've been doing, or for whom I've been doing it, but I am, > of course, always available for such work for any clients that believe > their > logging requirements are such that they have to eek out the maximum value > of > every last processor cycle in the software. > > > * FastFormat > FastFormat uses a similar technology to Pantheios to afford total > type-safety and unbeatable performance when formatting text statements. It > fully supports I18N/L10N by the use of numbered arguments; arguments used > more than once are converted only once. It supports arbitrary destinations > of the formatted result by the mechanism of "sinks". There are several > stock > sink types, including strings, string arrays, files, speech (currently > Windows only). > FF is almost ready to go - I've built the makefile template and proved > it on Linux/Mac/Windows with several compilers - and should be out next > week. Am just working through some other stuff first, and then plan to > write > the FF chapter for "Monolith" at the time that I release it (and submit > article(s) on it to mags). > Although my publicity machine has never been the best, I believe that > there should be no reason why FF should not become the de facto standard > for > C++ formatting/output, as it addresses all the problems of C's streams, > C++'s IOStreams, and the other open-source "modern" formatting libraries. > Time will tell ... ;-) > > > * flecxx > > One of the more simplistic messages in "Monolith" is that good > libraries > should afford one the ability to communicate with them in the types with > which you're writing your application code, rather than the > lower-level-of-abstraction types in which their interfaces are expressed. > flecxx, to be released in Jan/Feb, is a 100% header-only adaptation layer > library that wraps standard and popular 3rd-party libraries and removes > the > need to sully your beautiful code with all those .get(), .c_str(), > .GetSafeHWnd(), etc. calls. > > > * xTests > After releasing all these libraries, it's about time I put out the > testing framework I've been using these long years. This is particularly > so > because the next release of Pantheios, and every release of FastFormat and > flecxx, will come bundled with the testing framework. Open-sourced and > rebadged as xTests, it'll be available for anyone to use with any projects > sometime soon. > xTests is deliberately not all that sophisticated, so it's not going to > be a replacement for any xUnit-like library you're using. But it's simple, > and has minimal coupling, and does the job. > > > * xContract > Testing's important, but it's only half the software quality assurance > picture.Contracts are the other. I'm going to put my money where my mouth > is > this year, and reify all my wild theories about Irrecoverability in the > form > of the xContracts library. We'll start with C/C++, then Ruby, and then see > where life takes us after that. > > > > That's a brief summary of my non-commercial effort over the coming year. > All > those people who've been understandably frustrated about the lack of > doc/support with STLSoft should, by the end of '08, have all their issues > dequalmed. > > A happy new year to all! > > :-) > > Matt > > |
January 06, 2008 Re: Merry Christmas, and an STLSoft New Year! | ||||
---|---|---|---|---|
| ||||
Posted in reply to Matthew Wilson | A little late, but...
Happy holidays!
And congratulations, as others have said, on being an example to follow ;)
Best of luck for this year!
Pablo
Matthew Wilson wrote:
> First, I'd like to wish everyone a healthy and happy holiday.
>
>
> And I'd also like to let you know what you're in for from me with STLSoft
> (and related projects) in 2008. So, in no particular order:
>
> * STLSoft:
> 1. Website. This is finally going to get the professional treatment. It
> will have a redesign, be moved to its own server (so stlsoft.org is no
> longer synesis.com.au/software/stlsoft), have a blog, have a place for
> publishing small (or even large) articles, have tutorials, and a whole lot
> more. This should happen in Jan/Feb. Later in the year I may also provide
> commercial support facilities.
> 2. Documentation. The first phase will be to get the STLSoft
> distribution scripts updated and to start re-publishing the Doxygen
> generated docs. This should occur in Jan. In a couple more months' time,
> once the site re-org is done, I will be providing better and more extensive
> documentation, in concert with some other "senior" STLSoft fellows. More on
> this at another time ...
> 3. 1.10 will be released soon. It will contain several new components -
> including an uber-efficient properties-file class - and maybe even a couple
> of new projects (of which more later). One important thing is for me to
> address all the remaining items on the NG: there are currently 62 marked
> unread (i.e. left to process), some of which go back to 2005! Finally,
> there'll be a first tranch of some much needed rationalisation of
> internal/support files/classes/namespaces, to simplify and reduce
> compilation times.
> 4. Later in the year I plan, if I get time, to release STLSoft+, which
> is a commercial enhancement to STLSoft, supplementing various STLSoft
> sub-projects with substantial services.
> 5. Support for other new libraries, including xContract (see below)
>
>
> * Synesis:
> 1. The website's going to get a radical reworking (and reduction), to
> reflect the commercial emphases of the last few years: development process
> consulting, open-source customisation, network-related custom product
> development.
> 2. All the to-be-released soon (some as back as 2004!) will be removed.
> 3. The system tools will be updated (for Linux & Windows, 32 & 64-bit)
> 4. Some articles and/or blogs
>
>
> * "Breaking Up The Monolith: Advanced C++ Design without Compromise"
> (http://breakingupthemonolith.com/)
> This is my next book, and is about how to get everything you want from
> C++ without compromise, particularly between efficiency, expressiveness and
> robustness. It discusses the technologies (in particular the Shims concept
> and the Type Tunnel pattern) behind Pantheios and FastFormat, as well as
> other techniques for breaking up monolithic software, maximising design
> values, and so on. I've been writing it on and off for the last 15 months,
> but am about to put in a solid 10 days on it, which should get me to the 50%
> completion mark. The rest will be done over weekends over the following 3
> months. For anyone that's followed Imperfect C++ & Extended STL (vol 1)
> and/or the progression of the STLSoft, Pantheios, recls, FastFormat, flecxx,
> VOLE etc. libraries, this is the book that explains it all, including the
> Shims concept, the Handle::Ref+Wrapper and Type Tunnel patterns, the
> principles of Intersecting Conformance, Irrecoverability and Removeability,
> and more. Should be out mid/late 2008.
>
>
> * Pantheios
> As you may know, this is, IMO, the ultimate logging API library,
> affording the user complete robustness (incl. 100% type-safety) and
> unmatchable performance characteristics. It is designed to work above any
> other logging libraries, so you can have maximal performance combined with
> the configurability of, say, log4cxx.
> Pantheios has been in beta for over a year, but is now very close to a
> final 1.0 release. The remaining tasks are:
> 1. put in xTests (see below) unit and component tests into the build
> 2. implement buffering and file-rolling for be.file
> 3. website rework
> 4. project files for VC++ 2003/5/8 & Borland Turbo C++
> 5. update documentation
> 6. compatibility with VC++ "safe" compilation
> I've also been involved with some commercial customisations of Pantheios
> for companies here in Aus and in the US, who believe that Pantheios offers
> them a significant competitive advantage. Who am I to disagree? <g> I can't
> comment on what I've been doing, or for whom I've been doing it, but I am,
> of course, always available for such work for any clients that believe their
> logging requirements are such that they have to eek out the maximum value of
> every last processor cycle in the software.
>
>
> * FastFormat
> FastFormat uses a similar technology to Pantheios to afford total
> type-safety and unbeatable performance when formatting text statements. It
> fully supports I18N/L10N by the use of numbered arguments; arguments used
> more than once are converted only once. It supports arbitrary destinations
> of the formatted result by the mechanism of "sinks". There are several stock
> sink types, including strings, string arrays, files, speech (currently
> Windows only).
> FF is almost ready to go - I've built the makefile template and proved
> it on Linux/Mac/Windows with several compilers - and should be out next
> week. Am just working through some other stuff first, and then plan to write
> the FF chapter for "Monolith" at the time that I release it (and submit
> article(s) on it to mags).
> Although my publicity machine has never been the best, I believe that
> there should be no reason why FF should not become the de facto standard for
> C++ formatting/output, as it addresses all the problems of C's streams,
> C++'s IOStreams, and the other open-source "modern" formatting libraries.
> Time will tell ... ;-)
>
>
> * flecxx
>
> One of the more simplistic messages in "Monolith" is that good libraries
> should afford one the ability to communicate with them in the types with
> which you're writing your application code, rather than the
> lower-level-of-abstraction types in which their interfaces are expressed.
> flecxx, to be released in Jan/Feb, is a 100% header-only adaptation layer
> library that wraps standard and popular 3rd-party libraries and removes the
> need to sully your beautiful code with all those .get(), .c_str(),
> .GetSafeHWnd(), etc. calls.
>
>
> * xTests
> After releasing all these libraries, it's about time I put out the
> testing framework I've been using these long years. This is particularly so
> because the next release of Pantheios, and every release of FastFormat and
> flecxx, will come bundled with the testing framework. Open-sourced and
> rebadged as xTests, it'll be available for anyone to use with any projects
> sometime soon.
> xTests is deliberately not all that sophisticated, so it's not going to
> be a replacement for any xUnit-like library you're using. But it's simple,
> and has minimal coupling, and does the job.
>
>
> * xContract
> Testing's important, but it's only half the software quality assurance
> picture.Contracts are the other. I'm going to put my money where my mouth is
> this year, and reify all my wild theories about Irrecoverability in the form
> of the xContracts library. We'll start with C/C++, then Ruby, and then see
> where life takes us after that.
>
>
>
> That's a brief summary of my non-commercial effort over the coming year. All
> those people who've been understandably frustrated about the lack of
> doc/support with STLSoft should, by the end of '08, have all their issues
> dequalmed.
>
> A happy new year to all!
>
> :-)
>
> Matt
|
April 25, 2008 Re: Merry Christmas, and an STLSoft New Year! | ||||
---|---|---|---|---|
| ||||
Posted in reply to Cláudio Albuquerque | "Cláudio Albuquerque" <cláudio@nowhere.com> wrote in message news:fl6sq2$218f$1@digitalmars.com... > Hi Mathew, > > Beforehand happy holidays! > > Without any doubt I have to agree with Mr. Adi Shavit, I will be honest that the thing that amazes me the most is the level of "production" you present and sustain (synesis + book + stlsoft + personal life + ... ) my daily job alone "sucks" the life out of me. Thanks, mate. Although I have to say I feel like I'm walking through treacle most of the time. (For example, FastFormat has been largely complete for more than a year, but getting out the door has been the longest PITA. Soon, now, though <g>) > Anyway, could you give some detail regarding STLSoft+, since you detailed all the other stuff. This'll be a non-free (i.e. commercial offering, still likely 100% header-only, though). Don't expect to see this happening until Q4 08, however. |
Copyright © 1999-2021 by the D Language Foundation