Archive for February 2007

在公司會被人當成是 40 歲，在 IRC 有時會被人「前輩」「前輩」的喊，叫得周身不自在，我真的那麼老了嗎？

自從換走了 Rice University 的 Trackback Validator，用 Simple Trackback Validation Plugin 取而代之之後，終於有 trackback spam 逃過我的法眼。comment spam, pingback 和 trackback 三者就算 trackback 最難對付。

在 database 裏看過一下那無數的 spam，內容都是那麼老套，但一時興緻起來，數一下每個 spam 的 user-agent，有點出乎意料之外。將比較像一般瀏覽器的 user-agent 都過濾掉之後，剩下的統計如下：

有些實在沒法估計是亂來還是真的，特別是來自別的 blog 站 (例如 Movable Type) 的 trackback。但也有些 user agent 很明顯地告訴我們一些資料：

1. Net::Trackback 和 libwww-perl 一看就知是用 Perl 寫的。別的 spam 程式要用也寧願用像瀏覽器的 user agent，不會選這個。
2. Java/* 的不太肯定是否用 Java 寫的程式……
3. 最可疑的是 TrackBack/1.02。Google codesearch 這次幫上忙了。找一找 “TrackBack/\$”，符合條件的 source code 來自一套 worm (www.thierryb.net/site/IMG/zip/tb_avec_mail.zip)。

Basically, existing implementations of ABC music notation are almost usable. Indeed there are still some critical problems that prevent it from being production ready, but it’s already much better than GUIDO notation — well, to put it more accurately, it is the rendering that matters. One picture is better than a thousand words:

Read the rest of content »

現在大陸加建這麼多公廁，用這個來做標誌不是完全對題了麼？之前大陸還有大量的低級垃圾報紙將錯誤的內容隨處亂報，就只有新華社轉貼法制晚報這篇正確地分辨到原本貼標誌的人也只是轉貼，不是原創。

這些標誌和著稱「朋友說謊」 (friend lies) 的奧運吉祥物比較，藝術性實在高得多了，五粒低層次的小豆釘和這麼優雅的標誌完全不能相提並論。

Automattic 的人果然有兩下子。現在我才知道，Wordpress.com 也支援 LaTeX，不單是 post，連 comment 也開了。即是說打數學公式不需要自己起一個 blog 再裝 LatexRender plugin，到 wordpress.com 申請一個帳戶就可以自由發揮了！對我來說這個消息絕對可以列入 ‘Cool!’ 的行列。

如果連樂譜也可以 render 出來，那我自己的站大概可以關門大吉。我自己架 wordpress 的目的也不過是這兩樣東西而己。雖然未至於好像 mod_tex 那麼瘋狂，但將來預計 wordpress.com 還會有多點新搞作。既然 latex 能用，那麼除了數學公式以外，分子式和化學反應應該也不成問題。

Now GUIDO music notation is also supported in my (still) unreleased Wordpress plugin (subversion repository available though). It is basically a modified version of Chris Lamb’s FigureRender plugin, dropping latex support and adding support for other music notations; sadly it seems Chris has no time to continue working on it. Dropping latex support is due to my discussion with Steve Mayer about merging improved latex configuration back to his LatexRender plugin.

BTW, I start to recognize supporting GUIDO notation is not such a great thing. Before looking more deeply into GUIDO specification and putting it under practical rendering, everything looks wonderful to me — a web service called NoteServer, an open source implementation of parser library under SourceForge, plugin available for Sibelius and Finale… everything points to the conclusion that it is a greatly supported notation format.

But I didn’t notice other clues: this project is in articulo mortis. Not really dead (there is still a little CVS activity), but slow to a crawl. No news and no release since early 2004. So-called advanced notation spec never materialized. The source code doesn’t even compile on my Linux box (probably a bit better on Windows and OS X?). Last but not least, it doesn’t really live up to its word. See a real testing below.

The fragments below show how GUIDO and Mup renders the first 2 bars from Prokofiev 8^th Sonata 2^nd movement. First it’s GUIDO:

Now it’s mup’s turn. THIS is what people really should see.

No specific tuning for both of them, except adjusting font size for Mup (GUIDO doesn’t even allow that). No cross-stave barline, ugly slur by default (I don’t want to do any trial-and-error with bezier curve, not to mention NoteServer doesn’t conform to its own spec now), doesn’t support dynamics in between staves, doesn’t support more than one string on the same beat (notice GUIDO has eaten the dolce string). GUIDO is good for research interest, but in no way can it match publication quality.

Feb 22 edit: the GUIDO source fragment is available in comment.