sell the most and sell the best

Yahoo! gives up quest for search dominance

What do Yahoo, Apple and Ferrari have in common?


Dave Taylor gave us a view that the story isn't inevitably an ending to win or die. I agree to that. What I can't agree to, however, is that the analysis to compare Apple, Ferrari, and Yahoo!.

Apple and Ferrari are both differentiating them from their competitors. That's right. But is it true also for Yahoo!? That's a doubtful question. And the doubt doesn't come from what their strategy is, but from the internet business ecology.

Apple and Ferrari produce excellent products hard to imitate. The cost to manufacture similar computer or sport car is huge. It's unnecessary to mention the brand effect.

So how about Yahoo!? First, the cost to develop web sites similar to Yahoo!'s products is relatively low. That's because of the nature of the production in digital age. This is also a problem annoying digital content vendors, by the way.

We maybe all agree to the fact that Apple and Ferrari have excellent products. But how to define the word 'excellent' in internet age. Could it be said Google search product isn't good enough?

Of course Yahoo! doesn't necessary to be number 1 in search market in order to be useful and valuable to internet users. But with the shrink in market shares, Yahoo! inevitably can't be to hold its scale now and their influence anymore. Of course Yahoo! doesn't go to crumble if their rank fall down to number 2, 3 or even worse, but at that time, what's deal with Yahoo!?

Don't forget that the reason Ferrari can survive is they sell less but more expensive. And Apple almost went to fall away for a time. Apple comes to life and thrive because of its fine and popular product iPod. The strategy of iPod isn't the one "don't sell the most, sell the best"
, but "sell the most and the best!"
/>

Don't be a well-behaved baby

Google will fight Bush Administration demand for search records

Google, Yahoo, MSN Subpoenaed In Anti-Porn Effort

Don't be a well-behaved baby.

所謂的以文找文

蕃薯藤的新聞頻道最近採用了國內搜尋引擎業者宇匯知識科技的酷比「以文找文」技術。大約快一年前我得知國內有這一家公司，以及他們的核心產品「以文找文」技術。經過測試的結果，我認為這項技術並沒有特別的突破。基本上就是中文斷詞實作的產品化而已，將一篇中文文章加以斷詞，擷取出 frequent terms，然後用這些 frequent terms 當做搜尋關鍵字。其技術仍然架構在現今關鍵字搜尋上。

以這一篇女大學生卡奴悲歌 每月三萬元 上網找人包的新聞搜尋結果來看。蕃薯藤找出 16 篇相關文章。裡面有 5 篇跟卡奴有關，其它的包括：

大陸清大客座教授在台對博士班女伸鹹豬手
徐懷鈺撞傷人被控傷害　賠償4萬2千元和解
涉嫌以鑽子襲擊女學生怪客收押

等新聞。

從以上新聞的關鍵字看來，宇匯的「以文找文」技術從我們要求搜尋的文章中，擷取了卡奴、女學生等字眼，並進行關鍵字搜尋，得出以上搜尋結果。這種作法是不是能達到使用者的要求，其關鍵就在於演算法自動從文章挑選關鍵字的方法是不是精確。而從實驗中看起來，宇匯的作法似乎並不是那麼的成功。

這種 frequent terms 自動擷取的技術仍然有很多問題。一個 term 是不是重要，並不完全取決於 term 出現的次數與頻率。而即使像女學生這個在上文中出現多次的 term，卻不是該文的關鍵性概念。

draft of GPL v3

GPL v3 draft analysis
New open-source license targets DRM

GPL puts more strict limitations on DRM and patent usage.

Some countries have adopted laws prohibiting software that enables users to escape from Digital Restrictions Management. DRM is fundamentally incompatible with the purpose of the GPL, which is to protect users' freedom; therefore, the GPL ensures that the software it covers will neither be subject to, nor subject other works to, digital restrictions from which escape is forbidden.

That means DRM products don't use GPL licensed softwares anymore.

Moglen said that DRM technology, which places limits on how consumers can play movies, music or other digital content, is "fundamentally incompatible" with the principles of the Free Software Foundation.

The change will instantly affect the DRM inbuilt devices using open source softwares like TiVo.

All encryption and authorization codes of GPL v3 software need to be included in the product distribution.

Complete Corresponding Source Code also includes any encryption or authorization codes necessary to install and/or execute the source code of the work, perhaps modified by you, in the recommended or principal context of use, such that its functioning in all circumstances is identical to that of the work, except as altered by your modifications. It also includes any decryption codes necessary to access or unseal the work's output.

Podzinger

TechCrunch introduced Podzinger recently. It's a service you can harness to search thousands of podcasts. The ability to convert audio to text makes Podzinger enable to do exact search better than just searching description of podcasts. The charming feature of the service is that it allows users to listen just the audio snipper that contains the search terms. Michael Arrington said he encountered errors when trying this feature. I tried and it works well, but I don't know why it's. In search results, every result entry has some time marks and text showing where the your search terms occur in the podcast file. You can choose one of those marks to listen. Once you click those marks, you see the seeking indication showed besides the result entry. When it seeks the snipper it plays. Cool and useful service.