智能行驶

Spatial mapping 是 Google car 首先带到民车的，之前是 DARPA 在 tank 上面搞，tank 才是在人车稀少的野地测试，Google car 是在公路上跑有几年时间 0 事故，那时我还很呕说我光是 VR 模拟就频频当机，他们是怎么办到的？智能车开始出现事故是这一两年的事，看到有一辆直直往分隔岛冲撞去，很惊讶那么明显的障碍物怎么可能没侦测到？得知是把分隔岛另一边道路的右边地标线当成是自己车道的左边地标线，这不是早在世纪初还没有 spatial mapping 以前就开始在公路测试的平面影像辨识与移动追踪吗？现在不是已经多了 spatial mapping 吗？就算看错了地标线怎么会不知道前方有障碍物？火场寻径与道路寻径是通用同套，管它是火场还是人车都是禁止穿越过的障碍物，这还不需用到类神经网络，这是不经过思考几近反射的简单立即判断，除非是同时出现多个多向快速移动中的障碍物才伤脑筋。在 VR game 里都 run 到驾轻就熟很平常的动作，在 reality 意外是怎么发生的？确定不是传感器故障或处理器当机？

AI 寒冬？明明现在我就热到要吹冷气降温

从上世纪末到这世纪初科技界所谓的 AI 寒冬是指 NN 克服不了技术瓶颈表现差强人意沈寂到了近年才在技术上取得突破，而今年非技术人员所谓再一次 AI 寒冬算是属于达康类型的商界情景。当前并非所有 AI 都用 NN 为佳如同人脑亦是有些事用精准逻辑表现较好、有些事用模糊感觉表现较好而采另一套反而表现较差，早就说过要双管齐下同时进行而何者临场结果可接受即采何者，所以不要单方面看到某些时候 NN表 现不如预期就开始在说什么泡沫寒冬。自动驾驶开始出现事故当然是要去改进而非因噎废食，试问在 AI 以前有因为交通事故频传就都禁车而不改进车体的机构安全吗？从来想不劳而获的都是搞金融的，泡沫就只有股市商式炒作在泡沫，技术是不能不持续改进，为的是实用。

Dynamically Distributing Softwares

Services are not only on the servers but also on clients or routers; however, dynamically distributing softwares are much better.

扫码付

扫对方码付款要注意先对着要付的QRcode再开，切勿先开再去对，移动中万一扫到别的就误付出去了。至于出示码付款则切勿使用，一旦开啟，移动中任何方向都可扫你收你钱；出示码收款则无此顾虑，所以QRcode仅限收费方出示供扫。

Distributing Serializable Object

No files; only Serializable Objects. 

No services; only Distributing Serializable Objects. 

Distributed is static; distributing is dynamic. 

To access the local cache of the Distributing Serializable Object can update it on the original host directly without any other networking.

To localize the Distributing Serializable Object to let the cache of it not be swapped out can have the same effect with the block chain.

分辨机群中的无人机有人机

外观看不出机群当中谁是无人机谁是有人机时派无人机群去冲去逼就可以试出来，因为有人机无法做出一些无人机才做得到的动作。

Serializable Object Store

在多个近似性质的 Apps 间切换，每为一个服务又要再下载另外近似性质的 App 实在很麻烦！有人说那就用 Web 就不用下载 App 阿！还不是要切换 pages ？要的是一个画面里的互动就能处理全世界大大小小的事，那又不要一个 App 全交给一家做，要怎么办？就各家提供 serializable objects 不提供 Apps ，弄个 object store 。

☁️Cloud

☁️Cloud 的原始意义在上世纪下半是通讯各端只要确保信息传抵甭管中间如何传，这世纪元年进一步展至只要有脑用甭管位何。基于此，打从元年始我走的就是始于上世纪末的动态分布式运算操作系统让分布于网上的多脑用起来如同在使用同一大脑，数年后才发现有人走的是服务器代管路线，虽也实现了动态配置，但让人以为那就是 Cloud 不就同上世纪末让人以为 Web 就是互联网？现几大企所为 Cloud 还是仅称 Serverless 较为合适，如同 Fabless 的 IC 代工，其为代管，也就是说 server 、fab 实际上仍是存在，仍就是 client - server 。真正的 Cloud 是使多机用起来如一机的动态分布式操作环境，没有后台前台之观念。

☁ Could Cloud

So many people regard servers as Cloud. However, your handy device is one piece of the Cloud too no matter how tiny it is. So many tiny devices could be organized as a super computer by the Cloud operating system. 

☁

There is no remote storage.
There is only remote memory.

Each host swaps the senses between local memory and local storage automatically at background by themselves.

However, you also can regard there is no remote memory.

The distributed operating system swaps the senses between local memory and remote memory as cache automatically at background.

By this way, you can regard the whole network as one host.
Without this, it is not true Cloud.

Basic AI

No matter the Mathlab solves the formulas or the Prolog drives the rules or Siri or Cortana processes the natural language, they are functions processing the strings as parameters. What I want is the processor directly runs in formulas, rules, and sentences as programs but not only processes the string as text.

There are two ways to call the objects to process while they are touched on the screen. One spends time, and the other one spends space.

The time spent way for few objects is to call all objects one by one to calculate whether the touched points are within the scope of the objects projected on the screen flat and then do something each cycle.

The space spent way for lots of objects is to assign the object references together with the pixels while projecting the dimensional objects to the screen flat and later call the objects directly by the references at the positions when they are touched.

Augmented Reality

VR = R ➡ VR + VR objects = AR 

👽

👽：We had crashed the UFOs to you that year the first electronic computer made out by the human being on the surface of the Earth, why did you still not copy out the omni vehicle more than half a century passed?

The telescopic limbs hidden under the spherical surface

Perhaps there is someway to make the dynamic omni spherical shell no outer suspended ring needed by the telescopic limbs hidden under the spherical surface.

船体吃水少水阻少航速才快
但是相对浮力小可载重就小
左右多体船航行时气流由下穿过可产生浮力
但是平行面上承受水流的拉扯扭力
所以船底各体必须做成可各自旋转
但是又不能仅靠一转轴支撑
于是就演变成左右前后多可旋转分离式浮体
而多体船会摇晃得比单体船还厉害
所以层次球船舱是必要的了
以磁浮或滚珠弹簧间隔内外球舱

环状推进涡轮

传统涡轮推进器实心轴占据整管大部份空间，气流只穿过边边涡扇叶片的小部份空间，那是不是可改成中空环状让中轴空间变成内舱？中空环状涡轮套在滚球动态贝壳外边。

球狀外形回波

外表同样是球狀，一个回波讯号放大、一个回波讯号缩小，差别在哪里？

回波讯号放大者民用便于搜救；回波讯号缩小者军用便于匿踪。

一个反射回来、一个穿透过去，相同的外形，究竟是何致此异？

Without the multi-processing

非仅因轮流共用 pipeline 才需 multi-processing，轮流共用 I/O 亦需 multi-processing，所以不会因为从循序处理机进化到动态联结网络式图论机就没了 multi-processing，芯片内解放了 Von Neuman bottleneck 但芯片对外 I/O 仍是极为有限，旧 OS 型态仍然要多张 graph nets 轮流共用 I/O。仅一种情况可以完全拿掉 multi-processing，就是使用一种自动融合多个应用为单一系统的机制，此乃为何我一再提及鸟生蛋蛋生鸟的动态物件机制：要能自动从物件群中解析归纳产生共同类别以达应用合并整合为单一系统的效果。

连电源钮都无，当有无线电源输入且內建电池电量充足时即唤醒，当无无线电源输入时即睡眠，所以移动在外是靠无线电池贴近与否在控制开关。唤醒后可藉由互动令其睡眠而仅充电，要再唤醒就移离无线电源再次贴近。至于系统死机需强制关机重启咋办？移离无线电源再贴近先启动一个开端让使用者选择关机、睡眠、重启或进入系统。

但若环境持续供应远距无线电源咋办？改以微距RFID贴近与否控制开关。

现是自拍镜头置于屏幕边匡才可用贴纸遮住防骇客激活偷拍，以后摄像是in-cell在显象素后面利用视觉暂留瞬间息显摄像，贴纸何用？将通讯port做成实体硬件开关方为根本之道。

Patterns ↔ Senses

试想，若你的听觉受器是如视觉受器一般是阵列收讯，声讯对你来说是不是如同视讯画面一般？如此你会直觉两种收讯是不同的处里吗？是刚好你天生听觉受器就仅有两个单点，才会直觉视与听不同，在你进化前暂把声讯当成是1x1的2D画面串流，那直觉就会是一样的处理了。

All kinds of signals are patterns, the patterns are resolved to be the senses, and the senses are patterns too.

传统处理器在理解文字是一维循序串列处理，现在将每个字当做一个像素的值看待，不就是一张二维图像？那你觉得可不可以平行并列处理？整个理解过程是整面 operators（ 动词、介系词、连接词... ） 结合旁边的 operands（ 名词、冠词、形容词、副词... ）同时在发生，整个画面感觉就像粒子接触结合的化学反应一般组成了 Sense structure 以于脑内进行思维。

同理逆推，由 Sense structure 合成 pattern 输出是咋样的情形可想而知。

所有的硬件做到最后都是要可无限扩展；
所有的软件做到最后都是要可分散运行。

All of the hardwares done finally will be infinitely expandable;
all of the softwares done finally will be operated distributably.

看似不同的事其实是同一件事

我做的是 Dynamic Shell 的 strategic control，是立体全空间的寻径，非仅一机动单位的寻径还有多单位彼此之间的协调寻径。当 shells 在外头跑完停靠聚落即磊堆宛如建物，由一堆 shells 为移动式房间组成的建物，不是把车艇停在一区然后步行到另一区的房间而是车艇即房、房即车艇。每个贝壳进出聚落磊堆的时间不一，即使在休息中也自动挪移以让欲出之单位轮流转至外边，所以停靠聚落一样需要协调寻径，于是无论在外跑或停靠聚落都是相同的控制。

若也用这些 Dynamic Shells 来做物流配送，聚落磊堆时不就相当物流中心的载货移动机台？所以外面的配送和里面的配货有何不同？完全相同嘛！Shells 既是移货机台也是活动仓储，货在 shells 间移交时又相当于什么？自动贩卖机。而自动贩卖机内部的调货和外部的调货不又是相同？差别只在于大小而已。

于是总结：

A structural object is composed of the structural objects the same as itself, and the structural objects compose to be a larger structural object the same as themselves. It is the infinite space recursive structure.

The Augmented Reality eye device: angle @ angel

The augmented reality system does not only let us see what can not be seen in the reality on our glasses but also can let us do not see what should be seen in the reality, so it is necessary to switch between the reality and the augmented reality quickly at any time.

Augmented Reality Operating System without editors by the Sense

What in your memory are senses but not text. Every sense links its symbol graph and presents the symbols temporarily. The symbol graph is a sense graph too.

Start from the space object in the sense graph to travel out the senses (presented by the objects in the augmented reality too) to compose the sentence to have the conversation with the space object. By this way to define or modify the object without editing. What needs the brain storm is the structure of the sense graph.