Workspaces are terrible as cui components.
They make double versions of ACAD toolbars instead of checking to see if they are on or off. In other words the cui didn't fix the doubled bread crumbs everywhere
Also the cui stops updating after a few changes even if it did the change it won't admit it until I close it and reopen it wondering how soon my retirement is coming and dusting off the cobwebs when it finally opens
Yeah, the CUI in 2006 was the main reason I loaded up 2007 as soon as I got it. I was immediately overjoyed that it was now much easier to use the CUI, and how much better it works, even though it still has major issues. (My initial joy initially turned to disgust, though, for a variety of reasons that are documented in various posts on this site and on AUGI. Happily, things improved dramatically with SP1, and I now would not use any version except for 2007.)
I'm not really sure what you mean about the Workspaces making double versions of ACAD toolbars... Workspaces contain nothing but locations of toolbars, and the workspaces are generally not created in the CUI (although you need to go into the CUI to edit the menus in the workspace). Are you talking about the mess that results from trying to copy a workspace between CUIs on the Transfer tab?
I have to sympathize with any developer.
User friendly source code in today's market is a security risk.
That is not true at all. Secure, well-designed, user-friendly code is very possible, just not widely created. I have been disgusted for years at the garbage most people create. But then, I spent years writing code in Objective-C using what used to be the NeXT software base, and the NeXT software (still alive today as Apple's WebObjects) was doing things in 1992 that are STILL not possible in Visual-C++ (and never will be). Most industry analysts state that probably 80% of all software projects fail. And only a very small fraction are completed on time, on budget, and with the full desired fuctionality. It's a sad commentary on the profession. Unfortunately, many of the decisions Microsoft has made have fostered this turn of events, which is one of the reasons Microsoft tends to put me into rants. But rants against Microsoft serve little purpose, so I've been trying to cut back...
Toomany jerks out there who buy the tools to walk out of the store and bash peoples heads in instead of trying to build something other than contempt.
I'm pretty heavily into Obejct-Oriented Design (OOD), and really enjoy programming without writing code. Unfortunately, no tool yet exists in public for this... (Rational has made efforts with their Rose product, but it's been hindered for years by their focus on C++...) In the late 90's, I co-wrote a very nifty prototype of a modelling tool/IDE, though, that worked as a front end for Apple's EOModeler. Rather amazing stuff. Object-oriented design of an application in a graphical interface (the ENTIRE application, not just the GUI or enterprise objects); persistent data storage of any object in any relational database available at the touch of a switch; support for any true object-oriented programming language (such as Obj-C, Java, C#, etc., but probably not C++, at least not without a ton of headache and reduced functionality); the ability to open up a window on the internal source code of any object at any time, for those rare times when it's necessary to actually
write code... Unfortunately, around the time the first prototype was starting to get almost usable, Apple dropped their support of Objective-C. I would really like to get that thing finished and out into the market. It would completely blow the minds of anyone who is used to the Visual-C++ and MFC (or even .NET) way of doing things. But now we'd have to rewrite it in Java in order to keep using WebObjects as the core. The Objective-C/EOModeler Open Source movement has been making progress, though, so maybe that's becoming an option...? In my mind, I envision a Johnny Mneumonic sort of programming environment, eventually with a VR interface where the objects look like umpteenth-generation Lego blocks, and programming is accomplished by grabbing objects from VR space and shoving them together... Now THAT'S programming!
That's all in the future, though, when everyday hardware can handle VR with ease; the version we were writing looked like a more standard GUI application. But the tool did more than just make development faster and easier. It also effectively enforced standards, and we were planning on ways to quickly create the 30 or so most common OO design patterns. If you look at any programming task in the right way, you can break the overall task down into standard pieces. Looked at the right way, there are probably only roughly 30 of these. ANY task can be broken down into these same pieces. It's kind of like DNA - there are only 4 base pairs, but they can be arranged in a huge variety of sequences, resulting in the complete panoply of life. And these 30 patterns are "tried and true". Using them not only makes development faster and easeir, it cuts way down on bugs. So by creating an OO modeling tool that supported round-trip engineering and design patterns, the tool would effectively "teach" good programming at the same time it facilitated it and served as an IDE, and it would help guide the developer into making wise choices. This sort of thing is something that very few people in the world are even thinking about, but hopefully this type of thinking will become more common in the future.
One big reason so many software projects fail, or the result is filled with so many bugs and problems, is that even today, most people who work in software don't know how to program, let alone create solid designs. C++ became the standard language of the industry, and C++ is terrible for OOD. It barely qualifies as an OOPL. So therefore most software engineers get stunted, and never really get fully into the OO world, despite what they may think. And then there's the "I can do it better than anyone" mentality, that results in so many people trying to reinvent the wheel, because they think no one else has ever done what they're doing. That is almost always incorrect. It's like someone with no civil engineering background trying to design and build a suspension bridge, but for some reason, when it comes to software people think everything is different. Another big issue is that there is still no centralized body in control of software standards, which is pretty pathetic. It's the reason why the .CSV file you create in Excel may not be parsed correctly in Autocad, as one example. This sort of thing should have been taken care of decades ago.
Maybe I am something of a snob. I don't know. But the software industry has an 80% failure rate, for no good reason! That gets me going sometimes, especially since so many people now think that it is "normal" and "to be expected". I know from experience that it doesn't have to be that way.
Plug and play became the plug and pay in a short time.
Security through obscurity is not a solution but a cop out and the business as well as the developer pays for it.
Constant security updates are necessary but just as malicious as the security threat exploited.
It's an endless circle and more or less I'm likely singing to the chior here.
Yes, I agree security through obscurity is worthless. That's why I believe Bush&Co are spouting drivel when they say their counter-terrorism techniques will only work if no one knows about them, and that reporters are working for the terrorists by publishing stories about what the government is doing.
The constant security updates should NOT be necessary. That's simply another example of the low standards that Microsoft and companies like it have trained users to expect. The reason for it is the fundamental flaws in the design of Windows. Sure, some of it comes from the fact that Windows grew out of a single-user OS with limited capabilities, but that doesn't change the fact that Windows is fundamentally flawed. It would need to be redesigned from the kernel up to really fix it. From what I hear, Microsoft is not even doing this in the next version of Windows - instead, they decided to make Vista secure by just turning everything off by default. That is not security, that is denial. Well-maintained Unix systems have been basically totally secure for years. The only way to break into them is to gain access in some non-software way, such as by stealing or guessing passwords, or gaining physical access to the machine. Unfortunately, not many Unix sysadmins know Unix well enough to keep all the holes closed.
But shoot... I should really be working on getting something like that CUI guide put together, rather than ranting about this stuff. It serves more purpose.