原帖由 小牛军队 于 2007-2-19 18:53 发表
首先说说测试问题。我猜测，你说的一个星期的测试是对成品的测试。且不说一个星期的成品测试期对于交付到客户手中的软件来说是百分百的太短，单纯从你说的一个星期测试我就可以提出几个问题：你们总共测试了 ...
i think the Programmer who working in Microsoft is the luckyest one of the whole world, because they "make no BUGs", and every their Problem in Customer, it can be said just Customer's fault, not Windows. :D :D ,
the only question is then ,why they give so much service pack 1,2,3,4,5,6...one after another, since their code are BUG free? on other words, why they didn't "test" it before they sale it to customer?
i trust the person in our Software Development Devison much more than the guys in Microsoft. they have passed CMM III for a quite long time, and as i know they wrote even test program for Function Block over 2000 Lines.
1 Week test is passed for fast all projects, there is a "Badwell" prinzip, in China's market fast nobodyuse "monte carlo" prinzip for test. and except that error we have keep 5 Years BUG free record in Customer side. it says lots than the theorie from Prof. in Uni.
the real problem of Windows is, they offer only a set of function's interface, and a few Documents about the errors, memory control process,etc. and if a error hapens, which the Windows return Error you didn't know,then there is no why to find out the problem and solve it. But with Linux, the open sourcce systems, we can manually Debug and compiling the source what we need. for Windows, the only thing we can do is call and wait.
fast every service which Windows offers, can Linux also. for example Photoshop, there is a open source software called GMIP, fast the same function,and the reason why we need to use Windows, such a "good"system, is just because we used to use illigeally software like photoshop for a really long time,and if you use linux for a period time, you will forget windows.
use Windows, perhaps you will fill more easier, but you have to suffer from it's unstable, illogical errors, it's "stupid", until some day you have to reinstall it. and again and again, why it is unstable? because they have the "Best" employeer, they write BUG free Windows, and with Service Pack.
就我所见，工业领域使用少量的UNIX，大多数用Windows，基本上不可能用Linux，因为Linux下面根本没有工业用的软件。
原帖由 eisenstange 于 2007-2-21 13:53 发表
i think the Programmer who working in Microsoft is the luckyest one of the whole world, because they "make no BUGs", and every their Problem in Customer, it can be said just Custo ...
也许你没做过产品开发吧。unit test是最没意思的，但是也是必须走的。2000rows就我2个星期的代码的测试行。多层结构我不可能一起调试，只有每个重要的class都做unit test。 我真的很难想象世界上有只需要1个星期测试就那么健壮的产品。 也许环境变量为1吧。或者是单细胞的console programm。
那这根本和windows还是unix没有关系。
$bye$ $bye$
原帖由 eisenstange 于 2007-2-21 13:53 发表
i think the Programmer who working in Microsoft is the luckyest one of the whole world, because they "make no BUGs", and every their Problem in Customer, it can be said just Customer's fault, not Windows. ,
the only question is then ,why they give so much service pack 1,2,3,4,5,6...one after another, since their code are BUG free? on other words, why they didn't "test" it before they sale it to customer?
虽然你是E-Techniker，但我还是建议你去稍微听一听有关软件开发的课。首先我可以告诉你，测试的目标并不是你所谓的Bug Free，至于究竟是什么，留给你自己去寻找答案。。。你的原话我稍微修改几个词，你看看是不是也可以？
i think the Programmer who working in your Software Development Devision is the luckyest one of the whole world, because they "make no BUGs", and every their Problem in Customer, it can be said just Microsoft's fault, not their. ,
如果你们的软件在世界上拥有和微软一样多的客户，我几乎可以保证，你们出Service Pack的频率还要高。噢不，或者你们根本不会出Service Pack，因为你们连错误的根源都找不到？。。。
原帖由 eisenstange 于 2007-2-21 13:53 发表
i trust the person in our Software Development Devison much more than the guys in Microsoft. they have passed CMM III for a quite long time, and as i know they wrote even test program for Function Block over 2000 Lines.
1 Week test is passed for fast all projects, there is a "Badwell" prinzip, in China's market fast nobodyuse "monte carlo" prinzip for test. and except that error we have keep 5 Years BUG free record in Customer side. it says lots than the theorie from Prof. in Uni.
这是问题症结所在，你过度相信你们开发小组的人了，以至于连你们学校的教授你都不屑了。当然，相信同伴是一种美德，但是也要保持理性。和前面一样，给你提几个问题，让你去寻找答案：
1. 2000行的测试代码是什么概念？。。。噢，才看见我楼上的已经给了你一点小提示了。。。
2. 看来你是把monte carlo理解为测试的最佳方法了，这个观点可以说非常新奇。既然你不相信你们学校的教授，那么我建议你去好好阅读一些有关Test的书籍，看看Test究竟是什么。。。如果你连这个都拒绝去做的话，那我也就无话可说了。。。
原帖由 eisenstange 于 2007-2-21 13:53 发表
the real problem of Windows is, they offer only a set of function's interface, and a few Documents about the errors, memory control process,etc. and if a error hapens, which the Windows return Error you didn't know,then there is no why to find out the problem and solve it. But with Linux, the open sourcce systems, we can manually Debug and compiling the source what we need. for Windows, the only thing we can do is call and wait.
The real problem is, 你并不了解Windows，但这并不是大问题，要命的是，你不打算甚至排斥去了解它。微软提供的资料并不少，而是太多了，尤其是TechNet上面和MSDN上面，你觉得少的可能性只有一个：你访问MSDN和微软TechNet的次数比上萍聚少得多吧？。。。;)
原帖由 eisenstange 于 2007-2-21 13:53 发表
fast every service which Windows offers, can Linux also. for example Photoshop, there is a open source software called GMIP, fast the same function,and the reason why we need to use Windows, such a "good"system, is just because we used to use illigeally software like photoshop for a really long time,and if you use linux for a period time, you will forget windows.
这话很熟耳，仔细想想，原来是那个辍学清华博士曾经发表过的言论。。。
原帖由 eisenstange 于 2007-2-21 13:53 发表
use Windows, perhaps you will fill more easier, but you have to suffer from it's unstable, illogical errors, it's "stupid", until some day you have to reinstall it. and again and again, why it is unstable? because they have the "Best" employeer, they write BUG free Windows, and with Service Pack.
说句不好听的话，连windows都用不好的人只有两种：
1. 具有Linux宗教信仰而无条件强烈排斥Windows的人，比如FSF的开山祖师爷Stallman。。。
1. 在linux下多半也就是在KDE或者GNOME底下开开关关窗口的人。。。
Windows不稳定？重装？嗬嗬，过去六年内我家里的两台Windows电脑就只分别重装过一次，而且都是因为硬件坏了。你呢？重装过几次Linux了？。。。;)
如果你还想继续说服我Linux比Windows好的话，那还是停下来吧。我并不是*nix的反对者，相反，我很乐意在合适的场合使用*nix。之所以跟你说这么多，是想让你明白：linux除了广为传唱的好处之外也有很多问题，Windows除了广为传唱的问题之外也有很多好处，在你竭力攻击一个东西之前，先问问自己，你是否真的了解它？如果不了解，那就打开google，好好搜寻一番。。。另外，俗话说三人行必有我师，永远不要看不起任何人，对于你来说就是：不要看不起微软，不要看不起你所谓的“理论”教授，当然，若干年后也不要来个180度大转变而开始看不起开源社区的人或者看不起你现在的同事。。。$ok$
原帖由 shfaye 于 2007-2-22 07:45 发表
也许你没做过产品开发吧。unit test是最没意思的，但是也是必须走的。2000rows就我2个星期的代码的测试行。多层结构我不可能一起调试，只有每个重要的class都做unit test。 我真的很难想象世界上有只需要1个 ...
perhaps you have misunderstand what i said, otherwise i have wrong interpretiert.
"they wrote even test program for Function Block over 2000 Lines."
what i want to say, is, they wrote even test program for the Function Blocks in source code, which biger als 2000 lines, not the test code self biger as 2000 lines.
after CMM III(actually during the CMMII), Motorola und IBM(China) have meet with the leader team of SDD(Software Development Devision) several times, finally Motorola kick off their own team and buy the Software from out SDD.
the Prof. in Uni, we must also see wether they actually have a wonderful experience in Industry, otherwise what they taught, will be old, for example a Prof which is 50 years old, and if he taught Operating System but not Unix, how can he to prove he is better als some of the student ?
and till now linux Systemi didn't give me the chance to reinstall it.and even if i am a biginer in Linux, if there is new functions, i can recompile the Kernal, and for Windows?
连windows都用不好的人只有两种: 1. 具有Linux宗教信仰而无条件强烈排斥Windows的人，比如FSF的开山祖师爷Stallman。。。
1. 在linux下多半也就是在KDE或者GNOME底下开开关关窗口的人。。。
and how about the other guys who have to reinstall windows serval time a year just because of the system defekt and virus? perhaps they can use Windows so good, that the reinstallation just like a habit.
and can you give a reason why people need to give respect to 微软.
acutally we can make and we do make BUG free Systems, only the Prof. or the Students who take the words form "Prof." as motto, don't believe it, because using Windows, they don't believe there is BUG free Code anymore.
原帖由 eisenstange 于 2007-2-22 13:02 发表
perhaps you have misunderstand what i said, otherwise i have wrong interpretiert.
"they wrote even test program for Function Block over 2000 Lines."
what i want to say, is, th ...
天拉。。。。。。:(
你不要吧“单细包”开发就当成所有的软件开发了。这是误区。单片机通常没有过多的环境变量需要考虑，出现bug的情况自然就少。但是当我们处在比如说一个中型b2b产品开发的环境下，那就有比较多的异常情况需要考虑，作为一个普通人的大脑，错误是肯定无法避免的，唯一减少错误的方法就是养好test的习惯。比如说经常用unit test，喜欢工具的话junit，cunit等。真正的bug free在软件行业是没有的。就好象你不干保证你写了1万字的文章不能保证检查一次就可以查出所有错别字和语法错误一样。这和在linux环境下或者windows环境下开发根本没有任何关系。别把开发平台和操作系统概念搞混淆了。
原帖由 eisenstange 于 2007-2-22 13:02 发表
...
acutally we can make and we do make BUG free Systems, only the Prof. or the Students who take the words form "Prof." as motto, don't believe it, because using Windows, they don't believe there is BUG free Code anymore.
看到这一句，我。。。好吧，用linux可以开发出bug free的产品。。。你赢了。。。:(
原帖由 shfaye 于 2007-2-22 17:12 发表
天拉。。。。。。:(
你不要吧“单细包”开发就当成所有的软件开发了。这是误区。单片机通常没有过多的环境变量需要考虑，出现bug的情况自然就少。但是当我们处在比如说一个中型b2b产品开发的环境下，那就 ...
终于可以输入中文了，呵呵，用单片机开发，俺还不会呢，但是从操作系统角度来说，一个好的操作系统，是联系应用程序和底层硬件的中间层，如果我们把每个计算机硬件的指令都看成是理想的，那么操作系统就应该是一个良好的命令解释器，将用户的指令转化为计算机的指令，那么通过比较能否很好的进行转化，我们就可以对操作系统的特性进行比较，同时，一个同样的命令翻译过程，操作系统必须要尽量少的使用计算机硬件的资源，而不是占用了大量的内存，却不干活，这是效率，同时系统的开放性，健壮性，可扩展性，以及易操作性，也同样是衡量一个操作系统优劣性的重要标准。因此我们完全可以基于上述的几个关键点，给Linux和Windows打个分。
就前几点（效率，开放性，健壮性，可扩展性）就不多说了，应该大家都清楚，而Windows在易操作性上，确实对于90年代的普通用户来说，是划时代的，因为那时的操作系统不是给个人用户开发的，普通用户并没有见过任何视觉上的操作系统，因此在Unix下，从1996年开始，也开发了KDE这样的图形界面方便用户，在KDE上可以实现几乎所有Windows下都能实现的功能，因此，既然有了这样一个选择的机会，那么我们就应该让大多数人有机会了解这样的一个操作系统，从而真正的认识什么是一个好的操作系统，完全替代Windows也不是没有可能的事情。
至于编写无BUG的软件的先决条件是，（至少从开发部门的同事的经验以及CMM过程中的理解），应该是公司员工的知识体系从上层的应用到底层的接口都必须完全清楚，当然完全看完并看懂整个Unix的同事，我想信没有，但是我知道至少有两到三个在网卡驱动部分和内核部分是非常精通的，通过合理的定义每个函数的功能模块，并统一规范代码，确实可以做到软件的无误差。因为所有的东西，只要真正逻辑化了，就不会有错误了，而Windwos提供的库函数，或者说它提供的功能，并非是完全“符合逻辑”的，因此在Windows上的误差是不可能避免的。
长知识了！$支持$ $支持$ $支持$
不过，好象跑题了吧$汗$ $汗$
确实长见识
68楼的见解可以归纳为“软件乌托邦”$汗$