Frankly Speaking Reader Comments

This discussion has been archived. No new comments can be posted.

May 8, 2000

O'reilly has always been pretty hit and miss. It's just when O'reilly hits, it's usually the best around*.

On one hand there's the indispensable "Perl Cookbook" by Tom Christiansen & Nathan Torkington then on the other hand there's the completely unreadable "Advanced Perl Programming." by Sriram Srinivasan.

No publication house can guarantee the quality of their authors. The house can help by providing the author the time, access and resources, by providing editors who actually know the topic and by providing outside reviewers to catch errors. However all these measures will not "fix" bad writing.

When the publisher ruins a perfectly good piece of writing is when I start to question the value of the house. For example:

I found Danny Goodman's coverage, depth and level of communication more than adequate in "Dynamic HTML, The definitive Reference." However the layout and design of the book completely defeated the use of the book as a reference.

The visual clues and divisions between major topics, say a core object, and components of that topic where practically indistinguishable. They provided zero context if one was flipping through trying to find this or that property.

Browsing through the book, you'll see that practically every page had a sub-heading of "align" or "ID", but no indication of what these attributes applied to. In some cases the name of the element that is able to be aligned or identified was thirteen pages back.

Designing a reference on several topics (HTML, DOM, CSS and JavaScript) is admittidly a challange. It seems a challange Nancy Priest was not up to.

* With a very few exceptions, for example I was surprised that Damian Conway's "Object Oriented Perl" was *not* an Nutshell book.

As far as Frank's question about the CDROM included with Programming Python - I think it is not needed and it just increases the cost. I've had no trouble getting errata and code examples from O'Reilly and the actual program from www.python.org.

You may wish to update your comment about when Programming Python will have a new edition. Python 2.0 is now called Python 3000 (silly, but I'm no one to complain), and there will be a 1.6 and a 1.7 before 3000 comes out - is PP still going to wait?

I generally like O'Reilly's books. When I was first learning about UNIX in the early 90's my experiences with these publications vacilated between empowering enlightenment and total confusion. In the years since, my interest has focused on Linux, Perl and other open source programming environments. The quality of the O'Reilly books on these subjects is, by-and-large, still high.

My biggest complaint these days is the lack of electronic books from O'Reilly. I currently own their Perl and Networking CD-ROM bookshelves. I use the information in these books regularly and it's nice to have them in a portable, CD based format for reference purposes. I'd like to see them publish their newer Perl books, as well as their C/C++ and Java books, in this format also.

I know O'Reilly has some of their Java books on-line for paying members. This is nice but doesn't serve me when I'm developing at my computer without a network connection. It seemed for a while that one could buy deluxe versions of the books that included the web content on a CD-ROM. This seemed like a good offer. However, I never saw these "deluxe" books on the bookshelf before they were listed as "out of print" on the O'Reilly web site.

I discovered O'Reilly books many years ago, when searching for good books on some Unix-related topics. They were excellent (especially the Perl books), wonderful print, so O'Reilly has remained in my memory as the first-choice for Unix books. Then I moved to Windows and so lost contact with them.

Recently, I have got interested in Linux so I almost automatically turned to O'Reilly - a quality Unix books publisher - for a book. But I can't help my disappointment (this post is part of it). As a Linux newbie, I bought "Learning GNU/Debian Linux." The book contains many errors, is imcomplete on several points, difficult to seach for a particular information, the attached CD-ROM does not contain several programs the book instructs to install and describes how to use. And, worst of all, numerous places with explanations as: 'option config - configures the system' on one side, and 'if you press $#*f.&-Ox then you get that' - without explanation what is the principle behind, to allow me to cope with real-life situations. (Since the book is published under the GNU license, I may provide detailed feedback to the author for the good of the community - there is only the problem of time).

At first, I thought I had a bad luck. At least, O'Reilly has so many books nowadays that a weaker one is possible. So I turned my attention to "Running Linux" to find only a bigger disappointment. I didn't buy it so I do not know whether there are many errors, just concluded that with such low info/page ratio this book is not worthy of my time. My faith in O'Reilly's quality was shaken. I didn't know how to express it, but I found more persons who think the same and I think it is very sad. The books are not worse than from other publishers, they are simply not good enough.

Maybe I will give another try to "Linux in a Nutshell", but O'Reilly isn't a first choice anymore.

Disclaimer: everything I wrote here regards the content, not the form. The form of O'Reilly books is still excellent.

I have bought over 50 O'Reilly titles, so I've a decent sense of your product offerings and development over the past 6 years.

I think it is interesting that the two titles singled out are not "new" titles, but new versions. For myself, I found "Running Linux" to be one of the best introductions to a Unix-style operating system I have seen. I have recommended the manual to several individuals looking for a well-grounded introduction to the command line world.

I have purchased "Apache, the Defenitve Guide" in its last two printings. No complaints there either.

Learning Perl/Tk, however, was a complete dud. My feedback on this point was never acknowledged either.

The main critique I have with O'Reilly is the fufillment process. Orders are routinely late, and new release dates are a fiction.

Hope you take everyone's comments as seriously as you took the original poster's. That would ensure that I buy another ORA 50 books in the next 5 years.

I will pick an O'Reilly book on a topic over a competitor on any given day. A few bad apples in the bunch will not alter that.

If you look at the percent of bad titles within a collection, the numbers on the O'Reilly side are fantastic. Where with most other vendors you will be hard pressed to find the good books in their libraries.

The Stay Flat Binding should come back, however. Even if you have to charge more. It is definitely worth it.

Other vendors please take note. People don't complain about your bindings because the heavy glue bindings make for better doorstops.

I certainly agree that the quality of O'Reilly books has fallen.

My first O'reilly book was The access one, which was a NO NONSENSE technical book, not so much on access, but simple databases and logical queries. Similarly, the Transact-sql book was quite possibly the best reference ever written... Period.

Lately, however, there seems to have been a trend to marketing a massive number of books that are not so much great references, but aimed to be grabbed off the bookstore shelves.

In fact, for my references now, i rarely buy the O'Reilly books, prefering to hack through it myself with documentation of an API (for Java) or K&R or Stroustrup's C++ book.

I really do long for the days of old when If I needed an end-all be-all reference guide, I wouldn't even have to look inside the book, the O'Reilly name was good enough.

I too found myself completely in agreement with David Brickner's comments on April 15. As O'Reilly becomes bigger, hires new editors, and publishes far more books, it's essential that every editor hold very close to his or her heart the high standards that have made O'Reilly books so useful. Our editors are the guardians of the quality of our books, and making sure that no lemons slip by is THE key accountability of their job, far more than the requirement to get out a certain number of books or to generate a certain amount of revenue.

Readers should be aware that this isn't easy. Some authors find us tyrannical and very hard to please, and sometimes it seems easier to throw in the towel and let an author off the hook than it is to try again and again to get it right. This is especially hard nowadays, when the topics we cover are no longer out of the mainstream. As a result, there's a lot more time pressure, with competitors nipping at our heels. (By way of comparison, we want through six authors and six years, before we finally found someone able to do a good job on a sendmail book. We just don't seem to have that luxury any more.)

However, I do want to let you all know that it is still our goal to produce books that capture the knowledge of people who really know what they are doing, making it clear and accessible to those who want to learn from them. We don't want to slide down the slippery slope of "time to market", where any book, even a lousy one, is thought to be better than none at all.

I also want to let you know that all of our editors have been following this thread and we are taking very much to heart the concerns that you have expressed. Editors whose books are on the list of those that proved disappointing are very well aware that they have to raise the bar higher, and that they have to go back to work to make the next edition of those books worthy of your regard.

A final comment or two about bindings, another frequent topic: if you have a book whose spine cracks or whose pages fall out, please don't hesitate to contact our customer service department (800-998-9938) to arrange for a replacement copy. Like all publishers, we subcontract out our printing and binding, and if the spines are breaking down or pages falling out after only a few weeks of use, we need to know about it. We want those defective copies back, so we can show them to the manufacturer and get the problem fixed!

I do wonder, though, whether at least some of the reported problems do come from a misunderstanding of the "lay flat" RepKover binding we use on many of the books. For example, when I see a comment like T. Schell's:

"the binding of Java in a nut had completely separated from the book, necessitating the pouring of glue down the spine just
so the cover would stay on"

might suggest a real problem, but it could also suggest that we need to make it clearer (perhaps on the copyright page?) how the binding works. The pages are glued to a strip of cloth that is then glued to the front and back covers, but NOT to the spine. This allows the book to lay flat, but does create that gap, so disconcerting to the uninitiated, between the pages and the cover stock of the spine. This, by the way, is very similar to the way that a hardcover book is bound.

At any rate, I don't want to diminish by this comment any real defects, but I do want to make sure that everyone understands how the RepKover binding works.

In closing, I want to thank you all for your comments. Your high expectations of us are a wonderful gift, since it helps us remember how important it is to you that we do a great job.

I think we all agree that O'Reilly books are about the best we can buy on a subject. The last post about ORA being in competition with itself was right on target.
Thus far, I have not seen any real show-stopping problems or errors in any of the books I own. I have Running Linux, 2nd edition and love it (although I'm jealous of my friend's 3rd ed with almost 100 extra pages!) and could not get through my job without Essential Windows NT Administration. I'm about 200pages into the new Active Directory book, and while it has all the technical info I want, I feel like it's "missing something" as well.

What I have come up with is humor. I remember the TCP/IP book have some bit of levity to it, and that's what made O'Reilly the best: unmatched technical information in a readable format. This is probably the most difficult thing for the authors to do, and surely the first thing to go when they are under deadline, but I hope to see a return to the lighthearted nature contrasting with the serious technical content of the books. Maybe I'll see this in Essential Windows 2000 Administration!!

I just finished reading the article and the comments and I am left with an observation: O'Reilly is not in competition with other technical book manufacturers, it is in competition with itself.

Technical users have come to respect O'Reilly so much, that when they need help on a subject, they first look for O'Reilly's coverage before they look elsewhere. In finding the ORA coverage, they look no further. Perhaps they peruse the TOC and the Index, but for the most part they just go ahead an purchase the book and think that this is the best book they can get on that subject.

Now, they are rethinking that. Not because the other books by other publishers have gotten so much better, it's because some of the O'Reilly books just haven't measured up to the quality, depth of coverage, whatever that mystical quality that first reading of Essential System Administration or DNS and Bind had.

Frank, ORA must hold each book up to the quality standard it has already set, not to the poor one the competitors set. To me that means no unecessary fluff, good pricing, narrow defined topics covered in depth.

So far, my biggest dissapointment has been MySQL and mSQL. I gave up, bought the New Riders book when it came out and haven't looked back. I don't know if it is good or not, I just know it wasn't what I needed.

Someone in an earlier comment had the excellent observation that maybe some of us feel a slippage in the quality of the ORA books because we are now nearly equals in skill level with the authors (not necessarily on that subject) and as such we have different expectations. If this is the case, and ORA needs to satisfy both mature and beginning audiences, then a real separation between a "Learning" guide and a "definitive" guide need to be made.

Personally, I have never had an ORA book come apart on me. However, I am always careful with books period. I love the repcovers, the feel of the paper, the type size, everything. I hate all these things in books from nearly every other publisher. Just look at the Resource books from MS Press. The Win 2000 Pro book is 1600 pages and the type face is huge and 1/4 of each page has no printing. This book is unweildy and I will never buy it.

Well, all of this has got me interested in a trip to the bookstore.

i still think that o'reilly titles are outstanding. however, i've indeed noticed a certain decline in quality.

as tons of posters before me have mentioned, usually i check if oreilly covers the topic, and if yes, the search is over. well, it was. lately i start to poke around other publishers stocks.

why?

because to my liking, ora books become more like "the rest". when you covered hardcore tech details in depth in your older titles, the newer ones seem to be just "yet another book on a topic" - missing the gory details as i miss the in your typical run-of-the-mill IT book.

examples of books that kinda disappointed me include:

* UML in a nutsheel
(this _really_ reads like a phonebook. and actually looks like copies of overhead-slides)

* Javascript Cookbook
(while the perl counterpart offers a wealth of in-depth information, this one left me with a sour taste in my mouth. just nothing new in there)

* Programming Perl/TK
(hey, reading the online docu of TK is almost as good)

and wait, something else. PRINTED VERSIONS OF AN API ARE SIMPLY A WASTE OF TIME! i don't shell out money for a book that just covers an API that squarly sits in my browser anyway, for free.


but: GREAT JOB. i most definitely love most of your titles, if not all. please do everything to ensure that the standard keeps that high. for the simple reason that i'm far to lazy to search for other books - checking oreilly should suffice. in the past, for now and in the future :-)

A solid example of an O'Reilly book that failed is the MySQL & mSQL book. About 1/4 of the book was spent just explaining how a relational database works, and what SQL is. That stuff should be written once and put in a "Learning SQL" book or something. I wanted a book that would tell me everything I needed to know about MySQL, which already has excellent online documentation. The MySQL book by New Riders publishers is excellent, and exactly what I wanted. Look at that for an example of what I wish the O'Reilly book had been.

I love many of the O'Reilly books, such as Unix Power Tools, Programming Perl, Essential System Administration, and JavaScript: The Definitive Guide. Please work on making all the books of the quality of those I mentioned, and learn from what occasional good comes from an outside publisher like New Riders.

I, too, have a shelf full of O'Reilly titles.

I used to say 'If O'Reilly publishes a title on the subject, it is the first (sometimes only) book to buy on the subject. That is not true anymore.

The hard to define notion of usefulness is at the heart of the issue here. There just seem to be more books that are not O'Reilly's that are more useful to me in my highly idiosyncratic way of using books. I suspect that I am not alone though, thus this thread.

I dunno - in this age of 'Everything in 21 Days, unleashed", Maybe the kind of book O'Reilly used to produce is not commercially viable.

I also have a shelf full of O'Reilly books. They're usually excellent. However, I have also noticed some errors that got by proofreading. I don't mind these in the text, nobody's perfect. However, in the examples, they're a serious problem. Please pay special attention to proofreading the examples and diagrams. Also please actually compile and execute all the example code.

Your books have been incredibly beneficial to the community. I think many of us would be willing to look over a randomly selected chapter, and contribute corrections.

Thanks,
Doke

Hi gang,
As a recent (last two years) purchaser of O'Reilly books, I have found them to be still the far superior product. 'Running Linux' has a few weak spots for a newbie. "Essential System Administration' is a bit dated, yet still a heavily thumbed reference for me. As a newbie to Solaris as well as Linux, the O'Reilly books have been the best I have found for detailed information.

The pocket references are tucked into my attache and go where I go constantly. Twice in the past three months they have saved the day.

The best suggestion I can offer is hire a few more proofreaders and keep the details high.

Take care and be safe
pfb

I think O'Reilly has done an amazing job of keeping us all up to speed. You've got to look at their products in the context of the rest of the industry, and how much can change in a technology between when they begin to write a book and when they ship it. No one else rides this edge better they they do.

Another thing that has made me very happy with O'Reilly recently is the Pocket Reference books. Brilliant! I own most of them, and they're like little works of art. There's something almost Tolkien-ish about how they open right to the exact bit of info I needed on the first flip!

-Nick

Just a week or so ago I was talking with a co-worker about how I felt that the quality of O'Reilly books was "not what it used to be".

Like several others have said, I can not pinpoint exactly what has changed. I find the newer titles to be less useful, less interesting (the way the book reads, not the topics), and less accurate than some of the older titles.

O'Reilly built a reputation having technical books written by experts but for everyone. I think some of the original standard of excellence has been lost as O'Reilly has moved into new areas. In general I don't feel "Windows" books are as good as the "UNIX" books.

I still purchase O'Reilly books and I still recommend them to others. I ask that you address this issue quickly to prevent damage to the O'Reilly reputation and so that professional will continue to have a trusted "one stop shopping" source technical books.

Thanks for your time.

I just want to second the emotion generally expressed here that there is *some* slippage going on.

Here's an intangible: the books feel longer. One of the things I love best about ORA titles is that they are not tomes. But lately, in spots, some of them read like phone books.

I write out of affection, owning 20 plus ORA titles myself, and looking forward to many more books and years of reading pleasure.

BTW - Would it really drive costs up to fix the books so they lie flat?

First off, I am pleased to see that O'Reilly not only is interested in this sort of feedback, they have the guts to make it public. This is refreshing.

However, I have to agree with many other commentators that O'Reilly is slipping. You have some real winners:

* Learning Perl & Programming Perl
* DNS and BIND
* Java Servlet Programming
* Java Threads
* Java in a Nutshell
* JavaScript: The Definitive Guide

But you have some real lemons:

* Apache: The Definitive Guide
* Practical C++ Programming

Some books are so out of date they are nearly useless from a practical standpoint:

* NFS and NIS
* System Performance Tuning
* Essential System Administration

I know this is lacking in details, but I would be glad to share specifics with O'Reilly if requested. I own over 20 over their books and still recommend them over the competition. When I look for a book on a topic, I first look to see if O'Reilly covers it. If they do, my search is over.

I will say this much, the Apache book is riddled with errors. I am pretty competetent at administering Apache and I was hoping for some useful information about tuning and scaling, but what I found was trivial information about compiling and a reference to the httpd.conf directives.

Let one thing be clear... I don't want to purchase a paper version of a man page or a JavaDoc or a vendor's README. Reprinting this is just plain filler paper, and my arms are tired enough already.

Java Foundation Classes in a nutshell, Java in a nutshell, java examples in a nutshell-- all works of reference that might require being left open on a desktop for hours if not days. Within two hours of purchase, the binding of Java in a nut had completely separated from the book, necessitating the pouring of glue down the spine just so the cover would stay on. I found six grammar/spelling errors that a good proof read could have spotted if the book had had a inal read through. Don't feed me any c**p about getting the best information out at the lowest price. I'll pay the extra four dollars if the book actually stays together. You've really left a fecal taste in my mouth.

Java Enterprise in a Nutshell contains a number of errors and the Errata page still says: "There are no confirmed errors at this time." That page has not been updated since November 22, 1999. I submitted an error in January and it still has not been reviewed. I sent several e-mail messages regarding this unacceptably long period and have not received a reply. In general, I am happy with O'Reilly books, but some aspects of this book are not up to the usual O'Reilly standard.

I concur with another writer who finds that too many O'Reilly books fall apart. I have had two books with the hard-glue spine break in half within a few weeks. Use the flexable spine on all books!

I've been very happy with the content of O'Reilly books. However, the binding of a couple of my most recent purchases have started to separate almost from day one.

These books get a lot of use around here, please make them so that they last.

Paolo Damiani

I have been very happy with the O'Reilly books I have purchased. However, I hope they will update some of their older titles, such as 'Essential System Administration' (published in 1995). A lot as happened in the Unix since this title was published.
My most recent purchase was 'Unix Backup & Recovery' and I have recommended it to all my SysAdmin friends I know.
I have also enjoyed all of the Oracle O'Reilly titles.
Thanks for this forumn.

I agree with the comments about typographical errors, (See the Llama books...) If I released production code that riddled with mistakes there would be failures so get some testing done. What happened to the T shirts as well ?

In the past I always thought of O'Reilly books as a standard in book quality. Managing UUCP and Usenet and all these really good titles where the best books around and there were nearly no typos in there.

This all changed when I read the 3rd edition of Running Linux. It is one of the worst books I ever read from O'Reilly. But not only that. I submitted many typos to be corrected.

I heard nothing from ora.com but I looked on the errata again. I read through it and noticed that my corrections where there. Ooooppps... There where other typos added to them. Should I write an errata for the errata? Not that good.

I also don't know yet why I haven't been notified when you added my corrections to the errata and I would be really happy if you cross-check your errata entries with the customer who notices them in the future.

I'm really happy with all other books but Running Linux had much more typos than I expect from a 3rd edition book. (Many O'Reilly books have nearly no typos in their 1st edition)

Another question. Is it possible for external people to help O'Reilly to correct typos in the drafts? How can someone who is interested in specific topics be registered to get drafts of the chapters to proofread or something like that.

I think this question is also really interesting for others. Let the community help ora.com to prouce better books for all of us.

"DNS and BIND" is one of the best computer books I've read. It introduced me to O'Reilly books. After reading it, when I wanted a computer or Internet book, O'Reilly was one of the first publishers I checked. But I've had too many disappointments since then. I still check O'Reilly books, but I'm very cautious about buying one.

I eagerly anticipated the first edition of "Virtial Private Networks." It was terrible.
It took you four versions of your Java Deluxe CD-ROM to get it working right, and the online versions were a mess. I gave up on "Advanced Perl Programming" back in 1998. Have you looked at the errata for it? http://www.oreilly.com/catalog/advperl/errata/
I needed some advanced Perl, and I didn't have time to become a guru, or debug your book.

I know some people at O'Reilly, and I've heard some pretty disheartening things, including that you had cut back on quality control in the editorial and production process to save money.

Frank Willison, you say, "I have no internal data, though, that indicates a decrease in quality in our offerings." Do you have any data that indicates no change, or that indicates improvement? Do you have solid data about your book quality?

I have purchased a number of O'Reilly books (a few more than I have read, so far). Most are for Oracle, a few are for Unix/Linux and HTML. In general, I have been very happy with the Oracle books (especially those by Steve Feuerstein) and the others are also generally quite good. Although it seems to be a bit of a hodge-podge of online documentation, Running Linux proved invaluable during my first Linux installation 3-4 years ago. As I turn my attention to Java, I'm having a hard time finding a comprehensive, objective language reference (even among the O'Reilly titles). Most Java books seem primarily geared towards novice programmers and selling Java hype. Maybe this is just a function of the boom in IT. The audience has been dumbed-down (percentage-wise) and publishers are catering more and more to that audience. However, I do believe it is possible to write books that appeal to both experienced veterans and intelligent novices, but such books are getting harder to find. I'm not sure any more that O'Reilly is an exception to this.

I personally own over 25 different O'Reilly books
too and almost un-conditionally recommend them to
others, but technical errors in a technical book
are almost in-excusable.

Recently I went to The Perl Cookbook (still a best
value buy) for a canned solution. Now I don't
know enough Perl to know if it was an error in the
book, or an error in the software, but it took me a long time to find out how to get it to work.

In the process, I went to the errata section at
ora.com for the book and was surprised to see
"hundreds" of corrections (especially for the code
examples). After 6 new printings (I think) I
feel a little ripped-off that my first printing
edition is now so horribly obsolete not because of
any technology change, but strictly due to errors!

I'm almost inclined to demand a "warrantee
replacement" due to manufacturing defects in the
original material.

Perhaps you should institute the penny ^ N-error
rule from D. Knuth's books. Take the fee out of
the editors salary. After only a few errors,
and paying out the penalty, I'm sure the quality
would go up tremendously. (I seem to recall
that Donald Knuth hasn't had to pay out more
that a few dollars in the 20 years that his
books have been on the market.)

I have purchased a great many books on topics of all kinds from O'Reilly. Books on Apache, Linux, Java, EJB (I'm buying release 2 of that one today!).

Maybe I have a different take on this quality thing than others, but I think it deserves attention.

As I have matured in my understanding and my skills, I see things in text form that just don't make sense. I am willing to venture that the quality of the O'Reily offerings may be similar to days gone by, but that MY abilities now approach those of the authors.

When I first began to buy books, I was a novice in many ways. My experience (I run a successful e-commerce/business application ASP) has changed my perspective from that of pupil-to-the-author to that of peer-with-the-author.

I imagine that buyers that are now novices would report the same level of satisfaction that I felt several years ago. As they grow in experience and skill, perhaps they will also "gripe" about the quality.

Perhaps those that now feel the quality is lacking should propose topics that they could author and submit a proposal to O'Reilly? Can there ever be enough books on EJB 1.1? No, not at least until EJB 2.0!

Technology seems to be changing every week. Keeping titles coming that remain current is a challenge I don't want to face, and I'm used to deadlines.

Keep up the good work. I look forward to the next generation of titles from O'Reilly.

Every publisher puts out some lemons. However, O'Reilly is still putting out in most cases not only books that outshine similar offerings from other publishers but also books that *have* no similar offerings.

The two recent examples that come to mind are _Writing Apache Modules_ and _Ethernet: The Definitive Guide_. For what it is, I don't believe ORA has have published a better book than _Writing Apache Modules_. It takes an arcane, difficult, but timely and important subject, and makes it approachable without dumbing it down. I just read _Ethernet_ and went from thinking I knew a lot about the technology to realizing that there is a tremendous amount I don't know and I can learn it all from reading this book. Now that I think about it, I'd throw _Internet Core Protocols_ into the same category.

FWIW, I agree that Apache: The Definitive Guide is weak. So is the VPN book, and a couple of the Java books didn't work for me. But ORA's record is so good that it's still worth waiting for their book on any given topic, and it's especially heartening to see that ORA is continuing to publish books consistent with their original mission of doing books on subjects no one else would touch.

... reason that they are the best!

I am reading books from O'Reilly for the simple

I have definately seen a marked decline in the
quality of O'Reilly books. I see some Windows
books with menu structures laid out.
I was looking at Linux books recently and
although O'Reilly are not unique in this
respect I'm not impressed by seeing reprinted
man pages. Classic books like "DNS and BIND"
and other O'Reilly Unix books never wasted
space on that. In fact that was what really
separated O'Reilly from the rest of the pack.
No nonsense books that had what you really
need to know.

Hmmm... I've been noticing a sort of a drop in quality lately; now the O'Reilly stuff is generally only a single order of magnitude better than everything else available.

I have to say that I think most of the ``problem books'' are revamped FAQs & HOWTOs. It's always seemed to me like O'Reilly was providing a service instead of a product when they publish a book like ``Running Linux''; I know I shouldn't admit it but I can't deal with man pages and READMEs--gimme a book every time. I'd've never learned Perl well if it hadn't been for the Camel, Ram & Owl.

The thought that O'Reilly's quality might really sink to the level of some (most) of the other publishers out there is ehough to give one the yammering heebie-jeebies, but I don't even have goose-pimples yet.

I personally own 17 O'Reilly books, and we have about 30 sitting around the office. They are the best books on the market, but the new ones are obviously not as good as the older ones. As an example I use Programming Perl on a daily basis, but the Learning Perl/Tk book was a huge disappointment. Even worse was MySQL & mSQL I feel dumber from reading it, it was really bad. I couldn't believe that O'Reilly would publish that, it was truely one of the worst things I've ever seen. The Python books are fairly new aren't they? I think they're alright, but I think MySQL & mSQL needs a fresh set of authors/editors and a second edition.

I think that overall O'Reilly books are of a very high standard - in general they are my first choice. However, some books lately seem not to cover the depth with which we usually associate O'Reilly books.

For instance, I've started to use Linux after a several-year gap since I last used a Unix variant, and I have to say that Running Linux has a number of significant flaws. As a superficial handy reference it is fine - i.e. it's a reasonable "Learning Linux" book, but for an advanced computer user coming to Linux it is very seriously off track - details which one needs to find (e.g. say security, why doesn't telnet work, configuration files, window manager setup, tc, etc) which are beyond the very basic simply aren't there, or if they are clearly presuppose an indoctrination into Linux terminology, rather than using "transferrable" terminology which simply serves to obfusticate. For instance, take symbolic linking - mentioned in the text (page 98) - but nowhere in the book does it describe the corresponding "ln" command - bah!

I suppose I expect O'Reilly books to divide into "Annotated References" such as the (excellent) Java Swing book, or comprehensive guidebooks for experienced technical folks such as the (wonderful) Java Servlets book, or introductions such as Learning Perl. Several titles seem misnomered - "Running Linux" should perhaps be "Learning Linux", "Java Distributed Computing" should be something live "An Overview of..." etc.

Or, to coin a phrase, in a nutshell, several titles have lost the technical detail and clarity which are less measurable than error rates, but which we've come to depend on O'Reilly for. Programming Internet Email is another coffee-table book masquerading as O'Reilly material.

Okay, some of these are older books, but I think perhaps there is a gap in filling in new areas with paper with sufficient technical depth and precision which has grown.

That said, O'Reilly is in general still my first choice, and a number of recent books have been excellent - it's just that there have been just a little bit too much vagueness papering over cracks of late.

Well, this is kind of a difficult subject. Certainly, by comparison to any other Technical Publisher (some of whom I've written and edited for), O'Reilly is among the best.

But I also have to agree that it seems that compared with the O'Reilly of a couple of years ago the quality isn't where it used to be. The O'Reilly books used to be things that when I got them, they went straight to the working area of my bookshelf. Now, as often as not, they sit with the titles that looked promising but weren't.

It may just be the titles I got, as well. However, the "PalmPilot Companion" and the "UML in a Nutshell" are two of the more recent titles I've picked up - and regretted it. I never return a book, because I feel someone can always use it - but these titles are just waiting for me to find someone I can give them away to without feeling bad about it.

As your article mentioned, the issue of overall quality is a subjective thing. But I have to agree with those who are stating that the luster just doesn't seem to be there the way it used to.

Well frank you found that one too, I just thought well the rest of the text out-weights the error, all in all

We have in our tech library the offending title, but it's as worn and abussed as the rest ( why this guys use them as a coffe mat @ 2am I will never know ), still I wonder if they will last longer in hardback ?

Any comments publishers?

All in all I think I practically learnt sendmail thanks to bryan and eric

Javascript is a constant refernce for all of us, the memory jogging facilities provided by matt, matthias and lar, are great, seems to have become a book that vanishes all the time, but it can soon be found if you scream " where's my man page"

Have to go because I need to find my bison, should think david will be a great help in getting my cluster in line,

Have to admit though, New Riders Linux Firewalls is worth looking @.

Keep it up guys, much enjoyed reading all round.

yours a.r.b.u.

I like the quality of O'Reilly books. Yes, there are errors, but no more perhaps than with other publishers. Ever since the days when O'Reilly published several books on the 'Pick' DBMS (Hey Frank ... any chance of revisiting that area? Lots of changes since then) I always look for an O'Reilly book first on any subject I need to know about.

O'Reilly is the best tech publisher, no questions: full coverage, difficult topics, dedicated authors. More error checking, however, would be nice (especially in the thick Oracle books (which I spend a *lot* of time with) - Performance & Tuning ed2 is full of parity misses).

K Vainstein

I've found mistakes in O'Reilly books, but I've generally found fewer mistakes in O'Reilly books than in those of other publishers, and O'Reilly is quite responsive to error reports. A mistake I found in one book and reported was fixed in a subsequent edition, and the web-based errata information is very useful. They're not perfect, but they're still better than most of the market.

I'm glad that O'Reilly are willing to acknowledge that there may be a problem, rather than simply ignoring it.

The only O'Reilly book that I have been very unhappy with has been "The JavaScript Application Cookbook" which was neither use nor ornament. I had been expecting a javascript version of the immensely useful "Perl Cookbook" but got a half-hearted and very poor effort. The Javascript application cookbook is the only O'Reilly title I have returned.

O'Reilly books are always my first choice when
buying IT books. My shelf meanwhile contains about 38 "big" books and close to my PC I have about 12 of the pocket references. If I need to learn about any subject in IT I just look if O'Reilly has a book about it and then I know that I will find there all what I want to know. Especially the "... in a Nutshell" books are very helpful and often used.

Personally I love O'Reilly books. Obviously some books are better than others but no publisher of computer books satisfies as well as O'Reilly.

UML in a Nutshell was a disappointment, but overall I think O'Reilly has done an exemplary job of keeping their quality up. I remember why I stick by O'Reilly whenever I open a book by any other publisher.

Virginia Roberts

Return to: Frankly Speaking