The idea of using immune system to treat cancer is not new, but stays in a concept level with little clinical progress for long. However, with the exciting clinical results of BMS’s ipilimumab in melanoma patients, immunotherapy receives more and more attention. In the 2012 American Society of Clinical Oncology (ASCO) annual conference, the Phase I/II results of BMS-936558 in metastatic Non-small-cell lung cancer (NSCLC) were presented, leading the immunotherapy into a new era. BMS-936558 is fully human mAb blocking the programmed death (PD-1) on activated T cells. The PD-1 receptor is an inhibitory T-cell receptor that limits the immune response after binding with its ligand PD-L1 and 2. By over-expressing PD-L1, tumor cells escape immune system monitor. In the Phase I/II trials, both efficacy and tolerability of the new anti-PD-1 mAb were tested. A total of 75 stage IV NSCLC patients who had received at least one systemic treatment were enrolled in the study. After 12 cycles of mAb treatments, 19 patients showed different level of response and the overall response rate is as high as 18%. More impressively, BMS-936558 achieved 36% overall response rate in squamous patients. Anti-PD-1 mAb has general acceptable toxicity, with grade 3-4 adverse events (AE) rate of 8%. The most common side effects include fatigue (17%) and nausea (11%). However, 2 patients died of treatment related pneumonitis. The results of PD-1 immunotherapy are one of the biggest achievements for NSCLC treatment in recent years, especially for the 2 nd and 3 rd line therapy. Traditional chemotherapy usually has a 5-7% response rate, which is 1/3 of what Anti-PD-1 achieved. This clearly shows the therapeutic potential of PD-1. Importantly, PD-1 seems to have better efficacy in squmaous patients whom have very limited drug choice. Meanwhile, PD-1 only has function when T cells are activated and in the inflammatory site. It has favorable side effects profile compared to CTLA-4 targets, which have much broader effects on T cell activity. The only concern physicians have are the 2 drug related peunomotitis death. It is still not clear that such severe toxicity is a general effects for all PD-1 target, or specific to Bristol’s antibody. BMY has already committed for further development and all these questions need to be answered in the follow-on trial. Roche and GSK also have their own anti-PD-1 products under early stage development for different indications. Roche’s “armed antibody” becomes the other shining star in this years’ ASCO. Trastuzumab-MCC-DM1 (T-DM1) is the antibody-drug conjugate consisting of Roche’s major product trastuzumab (Herceptin) linked to cytotoxin metransine (DM1). After binding with HER-2 receptors, T-DM1 is internalized and then delivers metansine specifically to tumor cells. During the plentary session on June 3rd, Dr. Blackwell reported the interim Phase III of T-DM1 in metastatic breast cancer patients. A total of 991 Her2+ patients who had been previously treated with Trastuzumab and Taxane were enrolled in the study. They received either T-DM1 or Tykerb/Xeloda (TX) combination therapy. T-DM1 successfully delayed tumor progression (PFS) by an average of 9.6 months, comparing to 6.4 month in TX combination. More importantly, T-DM1 significantly improves the overall survival rate. After 24 months of the study, 64% patients in T-DM1 arm remain alive, while the median survival of TX combination is only 23 months, which has already showed significant differences. On the safety side, T-DM1 is also well tolerated. The grade 3-5 side effects happen in 40.8% of patients, 17% lower compared to TX combination. This favorable safety profile may be due to the high targeted chemotherapy agent delivery. In the last a few years, there have been significant shifts towards next generation mAbs platform, antibody drug conjugates are one of them. Through collaboration and licensing, Roche, Abbott and Novartis all have their own technology platform and mAbs under clinical development. The success of T-DM1 will certainly become an “accelerator” to this technologies and we expect to see more clinical results in the near future.
发展中国家很多实验室普遍存在生物安全性方面的缺陷，这些缺陷难免会造成病原体泄漏等生物安全危机，其影响可能会是全球性的。有关专家呼吁WHO等世界性组织牵头制定有关实验室安全的全球性标准。请看今天《自然》杂志发表的新闻述评。 Nature | News Biosafety concerns for labs in the developing world As report reveals lax standards in Asia-Pacific, researchers debate how to enforce rules. Ewen Callaway 22 May 2012 Biocontainment labs across the Asia-Pacific region all too often fail to live up to the term. An inspection of dozens of labs has found that nearly one-third of the biosafety hoods intended to protect workers from deadly pathogens did not work properly — an offence for which a Western lab could be shut down. In one facility, only a shower curtain enclosed a table on which the brains of rabid dogs were routinely dissected. Such deficiencies are symptomatic of a biosafety crisis in many of the laboratories that diagnose and study infectious agents in developing countries, say biorisk experts who attended a meeting at London’s Chatham House on 17May, where the results of the inspection were presented. The weaknesses could have repercussions around the globe if pathogens were released. “The strength of a chain is based on its weakest link, and developing countries are the weakest link,” says Teck-Mean Chua, former president of the Asia-Pacific Biosafety Association based in Singapore, which co-sponsored the anonymized laboratory inspection. Complaints of inadequate lab protocol in developing countries may not surprise many biologists, but they are attracting attention as scientists and research agencies in the West place increasing emphasis on biosafety. Discussions at the meeting skirted around the controversies surrounding the publication of research on mammal-transmissible forms of the H5N1 influenza virus. However, attendees did talk about how measures to protect lab workers and contain pathogens would affect research on diseases such as flu. In most Western countries, rules on biosafety — meant to safeguard lab workers against infection — and biosecurity, which protects the general public, became much stricter after 2001, when anthrax attacks in the United States raised the spectre of bioterrorism using laboratory-prepared pathogens. But stringent biosafety and biosecurity rules are unworkable in many developing countries, where researchers often need to handle infectious agents such as anthrax and plague to protect public health, but lack the infrastructure of the West, says Nigel Lightfoot, an associate fellow at the Centre on Global Health Security at Chatham House, who chaired last week’s meeting. “When you don’t have any electricity, the answer is not to build a very high-security laboratory,” he says. “You’ve got to move away from the costly bells-and-whistles solutions to what is practical.” Speakers suggested solutions such as small biocontainment boxes, for example, and also pointed out that pathogens that are endemic in a particular region present a lower biosecurity risk there than in Western laboratories. “We cannot stop them from working on things they need to for the health of their countries,” says Tim Trevan, executive director of the International Council for the Life Sciences, a non-governmental organization based in Arlington, Virginia, which is interested in biosafety and biosecurity in the Middle East and Africa. Lightfoot believes that “you’re going to have dual standards” to cover different areas. But having two sets of lab rules may not sit well with either side, says Nicoletta Previsani, who heads the biorisk-management team at the World Health Organization (WHO) in Geneva, Switzerland. Scientists in developing countries may feel that they are being left with less-than-safe labs, whereas those in richer countries could feel overburdened by regulations that others don’t have to follow. Some meeting attendees, including Chua, called on the WHO, the Food and Agriculture Organization of the United Nations in Rome and the World Organisation for Animal Health in Paris to take the lead in establishing global standards for lab safety and security. But Previsani says that such organizations cannot tell their member states how to operate their labs. Lightfoot adds that the WHO lacks the money and staff to act as a regulator. In January, the agency issued a five-year plan on laboratory biorisk management, in which it emphasized that it would be better placed in an organizational role, coordinating activities between stakeholders (see go.nature.com/xd9vdj ). Lightfoot argues that networks of non-governmental organizations and biosafety bodies ought to press developing nations to institute better lab standards. Donors could also help by paying more attention to the long-term sustainability of labs that they help to establish, adds Toby Leslie, an epidemiologist at the London School of Hygiene and Tropical Medicine, who has trained lab workers in Afghanistan. Too often, he says, money is spent on infrastructure and equipment without considering whether the laboratories can be operated safely by knowledgeable staff for years to come. A planned national health laboratory in Kabul, for instance, will need long-term support for maintenance and training. “I can’t see a way that Afghanistan is going to be able to support it independently,” says Leslie. Nature 485 , Pages: 425 Date published: (24 May 2012) DOI: doi:10.1038/485425a http://www.nature.com/news/biosafety-concerns-for-labs-in-the-developing-world-1.10687
At the year-end, everyone is requiredtowrite an annual review report or to fill in the annualreview form.Allhave triedtolist the whole year'sachievements inthe most possibly beautiful language. Although busy all day but at the year-end, it seemsthereis nothingimportantto write or the visibility of what we have done isunobvious. On contrast, blogging on the sciencenet is quite different. In spitewe only spend a very small part of time on blogging here, the visibility is very high.That's the power of internet. In 2011I have writenabout 60 blogs with contents varying from translation to editing, from work to life,from seriouspaper writing and editing tolight-heartedscenery and jokes.Only onearticle was recommended by the sciencenet editor (to be recommmeded is not the purpose of blogging), andseveral articles were shared by other bloggers or webs, butthe blog has been read tens thousands of times. Blogging can beaway of relaxation fromfatiguing work, and it's also agood channel toshare our thought and experience with others. Usually the last dayof the year-end isthe time tomakeresolutions, so I also make a resolution on blogging sciencenet-- I'llcontinue blogging here in 2012, and writemost articles in English. If what I writecanprovide some help to the authors orthecolleages, then that's my wishes!
一个非常棒的各种descriptor的 评测 。全文转载如下：（美中不足的没有特征匹配的评测）(一句话总结，综合考虑速度和性能，ORB是最好的。） Introduction For this test i have written special test framework, which allows me to easily add the new kind of descriptors and test cases and generate report data in CSV-like format. Than i upload it in Google docs and create this awesome charts. Five quality and one performance test was done for each kind of descriptor. Test cases Rotation test - this test shows how the feature descriptor depends on feature orientation. Scaling test - this test shows how the feature descriptor depends on feature size. Blur test - this test shows how the feature descriptor is robust against blur. Lighting test - this test shows how the feature descriptor is robust against lighting. Pattern detection test – this test performs detection of planar object (image) on the real video. In contrast to the synthetic tests, this test gives a real picture of the overall stability of the particular descriptor. Performance test is a measurement of description extraction time. All quality tests works in similar way. Using a given source image we generate a synthetic test data: transformed images corresponding feature points. The transformation algorithm depends on the particular test. For the rotation test case, it’s the rotation of the source image around it’s center for 360 degrees, for scaling – it’s resizing of image from 0.25X to 2x size of original. Blur test uses gaussian blur with several steps and the lighting test changes the overall picture brightness. The pattern detection test deserves a special attention. This test is done on very complex and noisy video sequence . So it’s challenging task for any feature descriptor algorithm to demonstrate a good results in this test. The metric for all quality tests is the percent of correct matches between the source image and the transformed one. Since we use planar object, we can easily select the inliers from all matches using the homography estimation. I use OpenCV ’s function cvFindHomography for this. This metric gives very good and stable results. I do no outlier detection of matches before homography estimation because this will affect the results in unexpected way. The matching of descriptors is done via brute-force matching from the OpenCV . Rotation test In this test i obtain pretty expectable results, because all descriptors are rotation invariant expect the BRIEF . Slight changes in stability can be explained by the feature orientation calculation algorithm and descriptor nature. A detailed study of why the descriptor behaves exactly as it is, takes time and effort. It’s a topic for another article. Maybe later on…. Scaling test SURF and SIFT descriptors demonstrate us very good stability in this test because they do expensive keypoint size calculation. Other descriptors uses fixed-size descriptor and you can see what it leads to. Currently for LAZY descriptor i do not have separate LAZY feature detector (i use ORB detector for tests) but I’m thinking on lightweight feature detector with feature size calculation, because it’s a must-have feature. Actually, scale invariance is much more important rather than precise orientation calculation. Blur test In this test i tried to simulate the motion blur which can occurs if camera moves suddenly. All descriptors demonstrate good results in this test. By “good” I mean that the more blur size is applied the less percent of correct matches is obtained. Which is expected behavior. Lighting test In lighting test the transformed images differs only in overall image brightness. All kinds of descriptors works well in this case. The major reason is that all descriptors extracted normalized, e.g the norm_2 of the descriptor vector equals 1. This normalization makes descriptor invariant to brightness changes. Pattern detection on real video Detection of the object on real video is the most complex task since ground truth contains rotation, scaling and motion blur. Also other objects are also present. And finally, it’s not HD quality. These conditions are dictated by the actual conditions of application of computer vision. As you can see on diagram, the SIFT and SURF descriptors gives the best results, nevertheless they are far away from ideal, it’s quite enough for such challenging video. Unfortunately, scale-covariant descriptors show very bad results in this test because pattern image appears in 1:1 scale only at the beginning of the video (The “spike” near frame 20). On the rest of the video sequence target object moves from the camera back and scale-covariant descriptors can’t handle this situation. Performance summary This chart shows the extraction time for N features. I made Y-axis as logarithm scale to make it more readable. For all descriptor extraction algorithm the extraction time depends on number of features linearly. Local spikes is probably caused by some vector resizing or L2 cache misses. This performance test was done on Mac Book Pro 2.2 with Core 2 Duo 2.13 Ghz. Further works Add new quality test cases. One additional test i know for sure – affine transformations. Your ideas for other tests are welcome! Add new kind of descriptors. Definitely will add an A-SIFT implementation. Create an LAZY detector with feature size and orientation estimation. Improve the LAZY descriptor extraction procedure. Expect at least 20% performance gain. Generate matching video for each test to demonstrate the behavior of each descriptor algorithm.
转自： http://www.cvchina.info/2011/10/08/zz-feature-descriptor-comparison-report/ 一个非常棒的各种descriptor的 评测 。全文转载如下：（美中不足的没有特征匹配的评测） (一句话总结，综合考虑速度和性能，ORB是最好的。） Introduction For this test i have written special test framework, which allows me to easily add the new kind of descriptors and test cases and generate report data in CSV-like format. Than i upload it in Google docs and create this awesome charts. Five quality and one performance test was done for each kind of descriptor. Test cases Rotation test - this test shows how the feature descriptor depends on feature orientation. Scaling test - this test shows how the feature descriptor depends on feature size. Blur test - this test shows how the feature descriptor is robust against blur. Lighting test - this test shows how the feature descriptor is robust against lighting. Pattern detection test – this test performs detection of planar object (image) on the real video. In contrast to the synthetic tests, this test gives a real picture of the overall stability of the particular descriptor. Performance test is a measurement of description extraction time. All quality tests works in similar way. Using a given source image we generate a synthetic test data: transformed images corresponding feature points. The transformation algorithm depends on the particular test. For the rotation test case, it’s the rotation of the source image around it’s center for 360 degrees, for scaling – it’s resizing of image from 0.25X to 2x size of original. Blur test uses gaussian blur with several steps and the lighting test changes the overall picture brightness. The pattern detection test deserves a special attention. This test is done on very complex and noisy video sequence . So it’s challenging task for any feature descriptor algorithm to demonstrate a good results in this test. The metric for all quality tests is the percent of correct matches between the source image and the transformed one. Since we use planar object, we can easily select the inliers from all matches using the homography estimation. I use OpenCV ’s function cvFindHomography for this. This metric gives very good and stable results. I do no outlier detection of matches before homography estimation because this will affect the results in unexpected way. The matching of descriptors is done via brute-force matching from the OpenCV . Rotation test In this test i obtain pretty expectable results, because all descriptors are rotation invariant expect the BRIEF . Slight changes in stability can be explained by the feature orientation calculation algorithm and descriptor nature. A detailed study of why the descriptor behaves exactly as it is, takes time and effort. It’s a topic for another article. Maybe later on…. Scaling test SURF and SIFT descriptors demonstrate us very good stability in this test because they do expensive keypoint size calculation. Other descriptors uses fixed-size descriptor and you can see what it leads to. Currently for LAZY descriptor i do not have separate LAZY feature detector (i use ORB detector for tests) but I’m thinking on lightweight feature detector with feature size calculation, because it’s a must-have feature. Actually, scale invariance is much more important rather than precise orientation calculation. Blur test In this test i tried to simulate the motion blur which can occurs if camera moves suddenly. All descriptors demonstrate good results in this test. By “good” I mean that the more blur size is applied the less percent of correct matches is obtained. Which is expected behavior. Lighting test In lighting test the transformed images differs only in overall image brightness. All kinds of descriptors works well in this case. The major reason is that all descriptors extracted normalized, e.g the norm_2 of the descriptor vector equals 1. This normalization makes descriptor invariant to brightness changes. Pattern detection on real video Detection of the object on real video is the most complex task since ground truth contains rotation, scaling and motion blur. Also other objects are also present. And finally, it’s not HD quality. These conditions are dictated by the actual conditions of application of computer vision. As you can see on diagram, the SIFT and SURF descriptors gives the best results, nevertheless they are far away from ideal, it’s quite enough for such challenging video. Unfortunately, scale-covariant descriptors show very bad results in this test because pattern image appears in 1:1 scale only at the beginning of the video (The “spike” near frame 20). On the rest of the video sequence target object moves from the camera back and scale-covariant descriptors can’t handle this situation. Performance summary This chart shows the extraction time for N features. I made Y-axis as logarithm scale to make it more readable. For all descriptor extraction algorithm the extraction time depends on number of features linearly. Local spikes is probably caused by some vector resizing or L2 cache misses. This performance test was done on Mac Book Pro 2.2 with Core 2 Duo 2.13 Ghz. Further works Add new quality test cases. One additional test i know for sure – affine transformations. Your ideas for other tests are welcome! Add new kind of descriptors. Definitely will add an A-SIFT implementation. Create an LAZY detector with feature size and orientation estimation. Improve the LAZY descriptor extraction procedure. Expect at least 20% performance gain. Generate matching video for each test to demonstrate the behavior of each descriptor algorithm. 分类: 新闻 标签: BRIEF , feature decriptor , LAZY , opencv , orb , RIFF , SIFF , surf , 描述子
对于研究生无论是做论文还是今后到公司从事研发活动,成果的主要体现形式就是Scientific Report （科研报告)。科研报告的重要性不言而喻。那么怎么才能写出比较readable的研究报告呢? 下面的几点总结或许有用。一般的报告都有如下几个部分组成，我们一一道来。 1. Title of report, name of authors (略) 2. Abstract The abstract should be very short Explain in a few sentences what you did in glance, the most important thing is what do you learn or what new konwledge do you get from this research? 3. Introduction In the introduction you describe the present knowledge You explain what you are going to investigate to obtain new knowledge 4. Materials and Methods. Describe the materials (source, etc) Describe the methods (equipments, procedures etc.) Often you can refer to the literature, so you do not have to describe the method in detail. Only exceptions or new methods are montioned. 5. Results and Discussion This is the most important part of the report!! Results can be compiled in figures and tables But, ... always start with text describing your findings and subsequently refer to the figures and tables Calculations are not given in this part You should ask yourself the following questions? Did you expect these findings? Can you explain your findings? If something went wrong, is there and explanations? Is the result from analysis A consistant with the results from analysis B? Try to make links between results!!! 6. Reference and Annex References Refer to scientific literature. It is not a good practice to cite the literature from one or few research groups a lot!! (including yourself papers) Annex Contains large papers like chromatograms, spectra, calucations, etc, 上面只是总结了写好基本报告的几个要点，对于初学的菜鸟还是非常有效的，尤其是化学、生命科学的研究生同学。要写出一篇佳作来，还需要长时间的摸索，学习和实践。 Wageningen university and research center
详见网站，有非常多对研究生有用的资源，欢迎转载和分享。 http://www.writing.engr.psu.edu/ Writing Guidelines for Engineering and Science Students Student Resources: Writing Exercises Slides: Scientific Writing Slides: Scientific Presentations Instructor Resources: Slides: Technical Writing Slides: Technical Presentations Instructor Handouts Teaching Slide Design Editor: Penn State, Leonhard Center Contributors: Penn State Georgia Tech University of Illinois University of Texas Virginia Tech Film: These guidelines for engineering writing and scientific writing are designed to help students communicate their technical work. To that end, these guidelines contain advice, models, and exercises for common writing and speaking assignments in engineering and science. Introduction Assessing the Audience Selecting the Format Crafting the Style Presentations Design of Presentation Slides Design of Scientific Posters Correspondence Memo Format: html or pdf Sample Memo: html, or pdf Letter Format: html, or pdf Sample Letter: html, or pdf E-mail Format Résumés Formal Reports Laboratory Reports Design Reports Progress Reports Theses and Dissertations Other Documents Proposals Instructions Journal Articles Conclusion Appendix A: Sample Format Appendix B: Documenting Sources Glossary References Acknowledgments Originated: 09/1997 Last updated: 11/2010 http://writing.engr.psu.edu/ All materials at this web site are the intellectual property of the editors and authors. You are welcome to make hypertext links to any of the pages at this site as long as you give proper credit. Please direct comments, suggestions, or questions to the lead editor: Michael Alley, Penn State.
New study reveals Russia's struggle to retain last-centuryprominence in scientific research Thank you for your interest in the Global Research Reports from Thomson Reuters and Evidence. Per your request, we are writing to inform you that a new reporthas just been released . This latest report, The New Geography of Science: Research and Collaboration in Russia , found that after reaching a peak in 1994 of just over 29,000 papers, output in Russia declined over the next decade to reach a low of 22,000 in 2006. To learn more about the changing landscape and dynamics of scientific research in Russia, download the new report here . To view all reports in this series please visit ourGlobal Research Reports website .
不知为什么有些国人比较迷信兰德公司 （ Rand Corporation ）。 所以就有人喜欢编造兰德报告行骗。而上当者不在少数。 90 年代有要在台海危机中出动航母编队打击中国的所谓兰德报告。笔者曾批驳过。后来 《 Dire Strait? 》 出版，谣言不攻自破。 最近又有一份关于中国的 兰德报告在网上被一再转载 。这 次 Rand 反应比较快，已经在其网站上声明：（ http://www.rand.org/hot_topics/china/ ） A report on the Web site dwnews.com incorrectly attributes negative comments about the Chinese people and Chinese traditions to a report about the Chinese economy allegedly prepared by the RAND Center for Asia Pacific Policy. No such comments exist in any report prepared by RAND researchers and comments in the article do not accurately represent the views of RAND scholars. Part of the dwnews.com article appears to be based on testimony given in 2005 that can be found here: http://www.rand.org/ pubs/testimonies/CT244/ . We have asked the Web site to correct its error. 这个假兰德报告中的第一段及第二段的前几句基本还是来自上面 News Advisory 中提到的 2005 年 5 月 William H. Overholt 在 U.S.-China Economic and Security Review Commission 的证言（ Testimony ） : : 如果 20 世纪的中国是一个富裕和统一的国家，我们会有一个完全不同的第一次世界大战，我们就不会有第二次世界大战而是第二次欧洲大战。中国 能够阻止日本侵略或者打败日本。美国在这些冲突上的花费从根本意义上会减少很多，因为珍珠港事件不会发生。我们和整个世界，更不用说 10 亿中国人，一个多世纪以来，已经为中国的弱小付出了惨重的代价。世界需要一个健康的中国。 中国的需求对日本走出衰退起到了促进作用。日本状况给世界经济带来了风险。关于这一点，怎么说都不夸张。日本巨额的债务会产生多米诺骨牌效应，逐渐波及到全世界。在中国有力的帮助下，危险似乎已经过去。 第三段则被翻译得很糟糕。原文： A more serious policy problem is hyper-competition created by cheap financing in China. The irrationalities of the Chinese financial system mean that in key sectors like steel China builds too many factories, and props up too many moribund companies, causing massive overcapacity. In recent years Chinese financial vagaries have led to excessive construction and huge demand for steel, aluminum, cement and others. For a while this has buoyed the global steel industry, including ours. But it has also led to construction of so many steel factories in China that soon China will have half of all world capacity. That means overproduction and eventually a steel price bust. 被翻译成： 中国金融体系的不合理意味著中国建造了垂死企业，导致巨大的生产力过剩。近些年来，中国财政政策上的反复无常导致过度建造，对铁、铝、水泥 和其他原材料产生了巨大的需求。日本人和现在的中国人看上去似乎会买下世界上所有的东西，但是当你看到他们的财政状况的潜在问题时，你会发现一个黑洞。日 本人在 90 年代陷入了这样一个黑洞，至今还在努力地爬出来。中国人很多年后仍将会为目前这种无节制的狂热的购买行为感到心痛。 第四段把： In the medium term China faces daunting challenges. Its banks are the worst in the world that we know about. In each generation a population about the size of the United States will move from Chinas countryside to its cities. Each year 12-13 million new workers join the work force. The impact of productivity on employment in manufacturing is much more severe than in our country. All these people need jobs. For a considerable period Chinas high growth can be sustained, but only through heroic reform measures by Chinas leaders. If somehow China powers through these problems, by 2020 its aging population will have the worst ratio of workers to non-workers of any population in the world, including Japans. That is to say, without some miraculous new policies the Chinese economy may well hit a wall in that period. In 2020, they will still be a very poor country by our standards. Even if their success continues until then, they will not be taking over the world. 翻译成： 目前，中国面临著巨大的挑战。中国的银行是我们所知道的世界上最糟糕的银行。中国每一代，都有相当于美国规模的人口从农村涌入城市。每年，都有 1200-1300 万新工人加入就业大军。在制造业，生产力对就业的影响比我们国家要严重得多。到 2020 年，中国人口老龄化会使工作人口与不工作人口的比率成为世界上最糟糕的，比日本更甚。如果没有特效的新政策的话，中国的经济在那个时期就会狠狠地撞墙。到 2020 年，以我们的标准来看，它会是一个非常穷的国家。 尽管做了很多手脚，比如 well 被翻译成狠狠地、 will still be 被翻译成会是，但基本上还算得上是翻译。 可是后面大量对中国人民、中国传统非常负面（甚至恶毒）的描述内容则完全是无中生有。（或者可能是移花接木？但是笔者目前没有找到出处。）时间也从 2005 年乾坤大挪移到 2009 年。 分析一下花了这么多心思编造这样一份兰德报告的人的心理，应该是一件很有趣的事情。 David A. Shlapak, David T. Orletsky, and Barry A. Wilson: Dire Strait? Military Aspects of the China-Taiwan Confrontation and Options for U.S. Policy, RAND, 2000 （ http://www.rand.org/pubs/monograph_reports/2007/MR1217.pdf ） 匿名：《中国现状分析报告》兰德公司亚太政策中心 （比如可以在这里找到： http://www.oursteps.com.au/ss/?action-viewthread-tid-169777 ） William H. Overholt, China and Globalization, Testimony present to U.S.-China Economic and Security Review Commission, May 19, 2005 （ http://www.rand.org/pubs/testimonies/2005/RAND_CT244.pdf ）
标签: Report
