NumeRe / DevLog: Recent posts

Today we publish the new release candidate no. 1 of the version 1.1.2. We appended the changes log for reasons of completeness. A brief overview may be found in the following:

• switch is a new control flow environment
• Clusters are new data types, which may contain numerical values and strings in mixed order
• We created new string functions, which may determine the types of the contained characters
• We improved the editor: saving of bookmarks, multi-selection pasting, autocompletion of LaTeX commands, warnings about global variables
• The console may now use HOME, END and CTRL-LEFT, CTRL-RIGHT correctly
• Animations include now an additional rendering step, during which the animation is cached to run the animation more fluently and less resource-consuming afterwards
• We improved the debugger massively: it now features many standard functions like step over block, execute next instruction and leave current scope
• The inline flag of procedures was improved and allows real inlining, which may offer runtime improvements of a factor 20
• NumeRe now offers an integrated version control of files, which are stored in one of the default paths and which are edited in the editor
• The dialog command provides graphical dialogs to interact with the user
• The refactoring functions were improved and may now be restricted in the search range. Additionally, each symbol renaming is logged
• NumeRe may now calculate a dependency tree of a procedure (Tools menu)
• NumeRe offers a package creator, which may create an install script from a provided set of procedures (Tools menu)
• The documentation window was redesigned and features now an index tree on the left and a toolbar on the top, which provides the history functions and the possibility to print a documentation article
• Many fixes and improvements

Some weeks ago,

we had the discussion, why we're using Apache SVN. Many of our colleagues use git as version control system and but we do not. They like the powerful features of git, which may offer many freedoms in their usage. And this might be one of the major issues in the whole discussion about SVN and git: SVN seems to restrict the freedoms of their users. I do not want to argue against it, because it's true. However, it's not always an issue and I'd like to explain, why that might be the case.... read more

Is it possible

that I'm silly? What's the reason for me to spend most of my free time working on an open-source project like NumeRe? I'm not kidding---that's a serious question, because from an objective, financial viewpoint there's really no reason to do so. I mean, of course I'm learning interesting stuff during my coding sessions, but I also could have done so by programming some commercial application. And, in fact, if I would work for a company, which would sell self-coded applications, I would most probably not work on this project, although the reason is different: after 8 hours of programming, I'm simply exhausted and neither able to produce some reasonable code nor even think of possible solutions.... read more

I complain quite often

about another numerical framework. It defines a domain specific language, where many identifiers are overused depending on where they are used in the code. If the identifier is used at the beginning of a line, then it is most probably a flow control statement, but if used in an expression, it is either a function itself or a dimension constant. From a software architects viewpoint in my opinion that's bad design and should not be implemented.... read more

I think,

everyone of us has this spot in his or her code, which he or she does not like. It's ugly, poorly written code back from the days, when we started our applications and learned programming. Code from these days is often single-function code, with many duplicates. However, it's working code, of course, and thoroughly tested--which is the reason, why this code is still there.

I have this spot in NumeRe's code, too. And, to be honest, there are more than one spots. They are working and not that central code sections. However, each time, when I have to change something in these sections, I have to fight against my own sickness and the strong feeling, to hit CTRL-A and DEL afterwards. I have reworked some of these sections already (those were mostly more crucial parts), but refactoring is not the top priority in my daily work. I want to focus on new features, which enhance the functionality of NumeRe. And I don't want to look back into bad written code, but rather into newer, elegant solutions.... read more

Since 2013,

I'm working more or less alone in this project. Of course I'm using libraries such as MathGL, TinyXML-2, GSL and so on, but more or less the whole work apart from the functionality provided by libraries was done by me. You bet that this is a lot of work to be done and it's really time-consuming. Well, at least during the testing phase I'm supported by a lot of people, although it's not a real testing phase but more a production test.... read more

Some days ago

one of my colleagues asked me, if I can describe NumeRe by its central use case. I really was baffled by this question. How might I describe such a complex application by only one use case, even if it is the central one? Some of you might have experiences with UML and use case diagrams--most applications do not feature the one and central use case, do they?

I escaped the situation by talking about the complexity of NumeRe and that it's not possible to name its one and central use case. However, this answer was not really satisfying--neither for my colleague nor for me. Sadly, I was not able to derive a better answer to his question.... read more

Last night,

we read at the project promotion site of SourceForge that we're encouraged to use the blog available here at SourceForge to write about NumeRe, especially, if we were doing a new release, because that would bring our releases to the attention of the SourceForge community team.

Nice idea, isn't it?

Yes, this was our first thought. But how would it look like, if we were only posting updated during a release? Okay, there are the regular updates in SVN and the notification, when we work on tickets. However, a blog containing only post about releases doesn't look very active, does it? So, we thought about content which would fit into this blog and won't need much deep-thinking during creation.... read more