Plauger：CUJ停刊互联网是主凶

（CUJ停刊）确有此事，但其起因并非众人所称者。杀死CUJ的，半个世纪以来一直在慢慢消灭报纸，而且目前正在更快地消灭印刷版杂志——那就是一个垂死的市场。单说技术杂志，互联网是主凶。过去我习于每月购买Byte杂志，了解微型计算机类新闻。现在我从新闻组获取这些信息，然后再利用搜索引擎挖掘更多细节。我也曾把Byte杂志当作真实性检验的手段——如果一种软件或硬件产品没有连续三个月在Byte上登载广告，我就绝对不会买。现在我搜索Google，查找产品规格、评测和最低价。想来别人也会这样做。

杂志（和报纸）依靠广告收入生存。而且它们就像酒店——高成本、杠杆利润。如此，成功者愈成功，失败者加速死亡。有些则不上不下、逐月苦撑，有时赢利，有时小亏。CMP拥有一堆杂志，而CUJ则是其中之一。老板们可以让它亏一阵，期望之后能有所改善。然而，生或死的判断总难避免。

如果回头看，就会发现，自从互联网公司兴起，所有的杂志都衰落下去，直至今日仍旧如此。数年以来，CUJ篇幅都控制在48页左右，这是基于成本和售价的平衡考虑。这样还很难维持，况且它也并非首难者。

我认为技术内容不是问题。我们这些创造和打磨它们的人，有责任使之与广告内容分隔开来。至少，对于长期撰写和编辑文章的我，在和关心钱的人员讨论经济问题时，的确是这样想的。有些命好的杂志，如《大都会》，根本无需考虑编辑性内容；人们（多数是妇女和理发师）买这种杂志，就是为了从广告中了解时尚信息。有些技术杂志骨气尽失地哀求厂商赞助。CUJ没有走上这条路。二十五年以来，我们月复一月努力提供有价值的技术内容。当然，质量也有参差，但我要说，大多保持了较高水准。所以，内容不是杀死CUJ的凶手。

最后我要谈一下，CUJ是不是“煤矿中的金丝鸟”。CUJ停刊，是否预示了C和/或C++的消亡呢？不然。人们阅读杂志，多半是为了看到新东西。1980年代，当C标准被开发出来时，我们这些在X3J11委员会工作的人，没完没了地讨论和写文章。1989年ANSI C标准出现时，这些活动都停下了。CUJ提供了更为实用的文章，告诉人们怎么用C，而不是“以后会怎样”。

也正是在那时，C++渐渐成为瞩目中心。好几本杂志、每年数个研讨会，都在探讨标准C++的成长。C++标准尘埃落定后，这些讨论也没有了。（还有Java热潮、关于各种吹牛语言的讨论等，不过我不想离题太远了。）C Users Journal变成了C/C++ Users Journal，并又赢得了十年时间。

今天，C和C++都是主力语言。Barnes和Nobel老兄只给出了两套架子，整个社群贡献了最终的语言和工具。这是因为，人们以讨论和阅读杂志的同样理由购买图书——学习新玩意，查阅技术信息、寻找支持工具、高阶教材、业界标准的培训、职位广告等等。（注意，我没有把入门级计算机科学教材包括在内。学院派家伙离现实世界太远。）你会发现，C和C++一直位居巅峰。而且，除却Java“猪过蟒腹”的那五年痛苦日子，过去几十年以来，始终如此。

所以，别只因为CUJ的衰亡而断言C和C++将死。PC Magazine自创始以来，目前处于最弱的状态，但世界上却有数以十亿计的PC在运行，而且每年还新卖出几千万台。PC更像是一匹驽马，或一种商品。PC世界的新奇之处在于其作为工具的新用途。这就需要各种不同的内容覆盖面。但你却无法让一本月刊充斥（电脑屏幕的）墙纸样式。

原文：

True enough, but probably not for the reasons generally espoused. What killed CUJ is what's been killing newspapers slowly for the past half century, and what's killing off print magazines more rapidly today -- a dying market. In the case of technical magazines in particular, the internet is the single largest factor of late. I used to buy Byte Magazine every month to get the latest news on microcomputers. Now I get much of my news from newsgroups like this one, with all the followup detail I can eat from subsequent web searches. Byte also served as a reality check -- I'd never buy a hardware or software product until its ads appeared for three consecutive months in Byte. These days I do a Google search for specs, reviews, and best prices. And I am hardly alone in this new modus operandi.

Magazines (and newspapers) live and die on their advertising revenues. And they're like restaurants -- large fixed costs and tremendous leverage on profits. Thus, a successful magazine (or restaurant) can make a bundle, while a loser dies quickly. It's those in the middle that hang on month to month, sometimes making a profit, sometimes losing a bit. CMP owns a boatload of magazines, so CUJ was more like a single restaurant in a chain. The owners might let it lose money for a spell, in the hopes that things will improve. Sooner or later, however, each entity must justify its existence, or get shut down.

If you go back and look, you'll see that *all* magazines began slimming down late in the Dot Boom, and have kept doing so ever since. CUJ has been around 48 pages for years; and that's about as close to life support as you can get. But it is also hardly alone, and not the first to die.

I've said next to nothing about technical content, and that's intentional. Those of us who generate it, or whip it into shape, are primarily responsible for keeping the ads separated by a bit of text. At least that's how it feels to me, as long time writer and editor, when I discuss economics with the money people. Some lucky mags, like Cosmopolitan, need diddly squat for editorial content; people (mostly women and hairdressers) buy it for the style message conveyed by the ads. Some technical mags make no bones about soliciting free, or even subsidized, copy from vendors. CUJ pointedly did *not* walk that path. We tried, month in and month out for 25 years, to provide articles of genuine technical merit. The quality did vary, of course, as did the focus. But I have to say that it kept a pretty high average. That was not what killed CUJ.

The last consideration I'll cover is the extent to which CUJ is a canary in a coal mine. Does its demise presage the death of C and/or C++? Not really. People read magazines mostly to learn about *new* things. When the C Standard was being developed in the 1980s, many of us in X3J11 were continually pestered to give talks and write articles. Once ANSI C was approved in 1989, there was a marked fall off in such activity. CUJ shifted to more utilitarian articles about how to use C, less of "here's what's coming next."

But at just that time, C++ stepped into the limelight. Several magazines, and numerour conferences each year, discussed the hothouse growth of Standard C++. That too began to fade as the C++ Standard settled down. (There was also the Java boom, speaking of strongly hyped languages, but I don't want to drift too far off topic.) The C Users Journal became The C/C++ Users Journal and won another decade.

Today, both C and C++ are workhorse languages, with little in the way of sexy new development. My local Barnes and Noble devotes only two shelves to both languages, freely intermixed. There are whole racks devoted to the latest fad languages and tools. And that's because people buy books for the same reason they go to talks and read magazines -- to learn about the new stuff. For the established technology, look at support tools, advanced textbooks, industrial training, job ads, etc. (Note that I do *not* include introductory computer science texts. Academic fads are only loosely coupled to real world needs.) There you will find C and C++ consistently at the top of the heap. And, aside from a five-year bump as the Java pig passed through the python, it's been that way for the past couple of decades.

So don't write off C and C++ as dead because CUJ withered and died. PC magazine is the skinniest I've seen it since it was born, yet there are about a billion PCs at work throughout the world today, and new ones sold by the tens of millions every year. The PC is much more of a workhorse, a commodity. What's new and sexy in the PC world are its novel uses as an appliance. And that calls for a different kind of coverage. But you can't produce a monthly magazine consisting mostly of wallpaper samples.