Write structured or unstructured Linux applications ?
Unstructure programming is easiest in Linux applications. You just add plug-ins and add-ons and extensions on firefoe, Konqueror or even Dillo. You have the choice of putting programs on toolbars or taskbars. Unstructured just means not sequentially important in the steps you take to do a task.
Structured is also easy once you decided the functions you want in the programs. Then you use a gui interface for each function in firefox type of browser. Gui interfaces will popup or pulldown to a menu of structured functions. Usually in native language files, whether you compiled them from languages or even mixed scripts. Very small modules as interchangeable parts in binary lines of codes are used.
This type of programming is very often used in tiny Linux distros. Puppy had made people aware by an exercise on abiword. Very impressive in the click and run world of Linux modernization. However, a lot of time was wasted. If only they had spend the time doing it to firefox with all the codec capabilities. Austrumi used it in their ppconfig for dialup modem to connect to ISP. Very clever but could be polished a little more, except they hope to instruct students of Linux distros to learn structured click and run programming in Latvia.
All this makes Linux click and run programmers learn fast and produce usable Linux distros with some simple automation. Then these structured programs, however small, can be incorporated into large unstructured application programs on their toolbars and taskbars.
- Login or register to post comments
- Printer friendly version
- 1207 reads
Preach componentization, practice embedded system ?
Componentization is the rage, virtualization adds more componentization. Emulation is a waste of time. We found Microsoft XP with 12,000 components, and every one component is compiled and depended on one another. 9000 for drivers, and 3000 components for functions. They go haywirely complex, its easier just to install them all because Microsoft gives you a template. The software is incredibly bloated and security becomes vulnerable.
What a mess until you can strip off some of the dependency issues. So, we are looking at embedded systems, where you only do one application at anytime. This means most of the hardware peripherals can be shut down via USB hotplug, software is stand alone. Resources of cpu usage is at a minimum. Try your hand at multimedia displays and you know what I mean. Cache is insufficient, data compression codec is too inefficient, and we have stuttering playing back on our computer.
If we can cache the whole episode in drams, would we have stuttered?
We are now practicing stand alone approaches. Browser can do both server and desktop .net framework and packet. So, server and desktop can have tight coupling without any other software involved in the communications(no emulation, all native codes). Mysql coupled to Mysql on both server and desktop must be of the same version to have a tight coupling. You can see the advantage of embedded system and tight coupling. It prevents any other hacker software from being effective to try and jam in when the data and istructions are not comprehensive(same platform).
You can not break into a running software, that is efficient(zero wait state and special keyed format).