Paul: Thanks for your comments on the previous thread.
Using bstrlib for the last few days has turned a rather painful trip down memory lane (returning to C after almost 20 years, and being reminded of the facilities that are left as "exercises for the user") into a pretty smooth-sailing experience. What a relief to have sensible string functions at hand!
So I got to wondering what aspirations you have for this library? It seems so suited to the needs of many, yet I don't see many references to it on the web, and it took quite a long time for me to stumble on it. Are ways available for you to push it more into the lime light?
Graham
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Do I look like an advertising consultant to you? :) I have no idea how to get the word out. I post this in my signature on USENET and maintain some links from my website. But beyond that? I dunno what else I can do. If you had any suggestions, I would be happy to hear them.
Paul
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
:-) Well, I'm not an ad consultant either... but it seems like this ought to be getting talked about in open-source circles. I'll be passing this on to a couple of scientific open-source/community-developed projects that I'm involved in, for sure. Various groups in the scientific community continue to be heavy users of straight C, and could produce much better software if using strings wasn't so fraught.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Perhaps the reason bstring is not more widely used is:
1) The name "bstring" appears to clash with an older BSD "byte string" package.
This creates confusion/FUD.
2) bstring documentation is a simple ".txt" file, instead of a Unix man page or Doxygen based.
This makes it less accessible (in my circle) and suspect (even if unwarranted).
I really like what I see in bstring and can see past (1) and (2), but others may not.
Correcting them may increase traction, even though the issues may seem silly to you.
I'm planning to use bstring in my next project.
Question: Paul,
1) would you consider a name change?
2) would you be interested in a patch for doxygen support?
Regardless... Thanks for a great package!!!
Randy
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
1) There is a *lot* of code (much of it written by me :) ) that uses the bstring names. It would be a big deal to truly change it. Its possible, but some users may balk. But at the very least you need to propose a good alternative. The BSD functions I stumbled upon some time ago, and thought I had worked around any naming conflicts.
2) Doxygen, as far as I understand it, extracts function leader comments, parses structures and gives a text mapping of it all. I am not a doxygen user and its not my intention for the documentation to be limited to descriptions of functions and structures. The documentation discusses performance, safety, compares it with other libraries, talks about trade offs of the design and so on. I don't want to give that up. The whole thing comes with a unit test and its all open source -- that's my way of making it trustworthy. If you can make a Doxygen patch that retains all the substance of the current documentation and I can maintain it, then I would be happy to include such a patch.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I should clarify (1). I was not proposing a name change to the functions themselves. I rather like your names and since there is no name-collision I'd keep them as-is.
My comment was more of an observation regarding "brand name recognition". When I googled "bstring", I found man pages for the BSD utilities. This caused confusion, for this "consumer" at least. A unique and recognizable "brand name" might help in some small way. I'm not an advertising consultant either, so maybe I'm just off base.
Alternative names... hmmm maybe something as simple as "bbstring" or "stringb" or "pbstring" (Paul's Better String), or ... choose your favorite word play. This would leverage existing name recognition that you already have. Candidate names should be carefully vetted for uniqueness among other things (which I have not done). The "brand name" is often used to identify a family of related routines in man pages... like BSD's "man bstring(3)".
Regarding (2), programmers appreciate documentation that is easily accessed, which includes a consistent interface to it. Unix man pages and Gnu Info pages are two common standards. They provide one-stop shopping/searching for a wide variety of documentation. I personally hate Gnu Info pages, but there are tools that can derive man pages from Info source, for ole-timers like me.
Doxygen provides a richer notation for documentation than either man or info. That would be my recommendation, even for an ole-dog like me that is reluctant to learn new tricks. Doxygen can output man pages or several other formats (but not Gnu Info AFAIK). It is not "limited to descriptions of functions and structures". PBString would not "have to give up anything". You might even like the automatic indexing, syntax highlighting of examples, cross referencing via HTML hyperlinks, etc.
My apologies if any of my comments were offensive instead of helpful.
Randy
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I'm browsing on the web to found a thread safe alternative to std::string, and I've discovered your job.
I'm impressed by your work, which seems very good, you've done a deep analysis on existing library and strings problematic.
But, as strange as it seems, as Graham said, your library is not very familiar on the web, and I think it's sad.
Some suggestion, to perhaps improve the first impression of people which pass on you site :
- as Randy said a TXT documentation, is not very sexy, perhaps just the same in HTML with a bit of bold, would be more sexy (and perhaps even, more useful)
Doxygen could be a complement,
- Add a link to an HTML changelog with recent date
- ask current users to said they use it,
- have a small tutorial page to show how it's easy to use.
- If you have tested it on recent compiler, say it, it shows the development it's still active.
The problem, with good, stable library, it's there's not a lot of message talking about them, then you can think they are not used.
I've took a quick look on your code, it seems pretty well written.
I will test it a while, and probably mention it in an future Article on a french magazine.
My congratulation for your job, hopping it will be recognized by the community
Laurent
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Paul: Thanks for your comments on the previous thread.
Using bstrlib for the last few days has turned a rather painful trip down memory lane (returning to C after almost 20 years, and being reminded of the facilities that are left as "exercises for the user") into a pretty smooth-sailing experience. What a relief to have sensible string functions at hand!
So I got to wondering what aspirations you have for this library? It seems so suited to the needs of many, yet I don't see many references to it on the web, and it took quite a long time for me to stumble on it. Are ways available for you to push it more into the lime light?
Graham
Do I look like an advertising consultant to you? :) I have no idea how to get the word out. I post this in my signature on USENET and maintain some links from my website. But beyond that? I dunno what else I can do. If you had any suggestions, I would be happy to hear them.
Paul
:-) Well, I'm not an ad consultant either... but it seems like this ought to be getting talked about in open-source circles. I'll be passing this on to a couple of scientific open-source/community-developed projects that I'm involved in, for sure. Various groups in the scientific community continue to be heavy users of straight C, and could produce much better software if using strings wasn't so fraught.
One obvious way is for people to blog about it and link to it.
Perhaps the reason bstring is not more widely used is:
1) The name "bstring" appears to clash with an older BSD "byte string" package.
This creates confusion/FUD.
2) bstring documentation is a simple ".txt" file, instead of a Unix man page or Doxygen based.
This makes it less accessible (in my circle) and suspect (even if unwarranted).
I really like what I see in bstring and can see past (1) and (2), but others may not.
Correcting them may increase traction, even though the issues may seem silly to you.
I'm planning to use bstring in my next project.
Question: Paul,
1) would you consider a name change?
2) would you be interested in a patch for doxygen support?
Regardless... Thanks for a great package!!!
Randy
1) There is a *lot* of code (much of it written by me :) ) that uses the bstring names. It would be a big deal to truly change it. Its possible, but some users may balk. But at the very least you need to propose a good alternative. The BSD functions I stumbled upon some time ago, and thought I had worked around any naming conflicts.
2) Doxygen, as far as I understand it, extracts function leader comments, parses structures and gives a text mapping of it all. I am not a doxygen user and its not my intention for the documentation to be limited to descriptions of functions and structures. The documentation discusses performance, safety, compares it with other libraries, talks about trade offs of the design and so on. I don't want to give that up. The whole thing comes with a unit test and its all open source -- that's my way of making it trustworthy. If you can make a Doxygen patch that retains all the substance of the current documentation and I can maintain it, then I would be happy to include such a patch.
Thanks for the reply Paul. You have good points.
I should clarify (1). I was not proposing a name change to the functions themselves. I rather like your names and since there is no name-collision I'd keep them as-is.
My comment was more of an observation regarding "brand name recognition". When I googled "bstring", I found man pages for the BSD utilities. This caused confusion, for this "consumer" at least. A unique and recognizable "brand name" might help in some small way. I'm not an advertising consultant either, so maybe I'm just off base.
Alternative names... hmmm maybe something as simple as "bbstring" or "stringb" or "pbstring" (Paul's Better String), or ... choose your favorite word play. This would leverage existing name recognition that you already have. Candidate names should be carefully vetted for uniqueness among other things (which I have not done). The "brand name" is often used to identify a family of related routines in man pages... like BSD's "man bstring(3)".
Regarding (2), programmers appreciate documentation that is easily accessed, which includes a consistent interface to it. Unix man pages and Gnu Info pages are two common standards. They provide one-stop shopping/searching for a wide variety of documentation. I personally hate Gnu Info pages, but there are tools that can derive man pages from Info source, for ole-timers like me.
Doxygen provides a richer notation for documentation than either man or info. That would be my recommendation, even for an ole-dog like me that is reluctant to learn new tricks. Doxygen can output man pages or several other formats (but not Gnu Info AFAIK). It is not "limited to descriptions of functions and structures". PBString would not "have to give up anything". You might even like the automatic indexing, syntax highlighting of examples, cross referencing via HTML hyperlinks, etc.
My apologies if any of my comments were offensive instead of helpful.
Randy
Hello Paul,
I'm browsing on the web to found a thread safe alternative to std::string, and I've discovered your job.
I'm impressed by your work, which seems very good, you've done a deep analysis on existing library and strings problematic.
But, as strange as it seems, as Graham said, your library is not very familiar on the web, and I think it's sad.
Some suggestion, to perhaps improve the first impression of people which pass on you site :
- as Randy said a TXT documentation, is not very sexy, perhaps just the same in HTML with a bit of bold, would be more sexy (and perhaps even, more useful)
Doxygen could be a complement,
- Add a link to an HTML changelog with recent date
- ask current users to said they use it,
- have a small tutorial page to show how it's easy to use.
- If you have tested it on recent compiler, say it, it shows the development it's still active.
The problem, with good, stable library, it's there's not a lot of message talking about them, then you can think they are not used.
I've took a quick look on your code, it seems pretty well written.
I will test it a while, and probably mention it in an future Article on a french magazine.
My congratulation for your job, hopping it will be recognized by the community
Laurent
Hello Paul,
Attaching your bstring library to the Boost C++ libraries would get it out there.
http://www.boost.org/
Boost current has no enhanced string library.
Regards
Mike