This book did not favor any one Linux OS over another (good). Most Linux configuration was via commandline (which most will advocate in a Unix environment) and direct editing of conf files. Each chapter has a great balance of common practices, breaking down the common config parameters, then implementing. Overall, a great reference.

Working several years as a certified Windows Administrator, I find it strange to have a complete chapter soley dedicated to the topic of NetBIOS. Who uses NetBIOS? I looked further into the book, and the all SAMBA material seems to be a rehash older and overdocumented SAMBA 2.0 docs...

I looked at another section and it seems that Windows NT 4 time was covered (maybe accidently through author's misunderstanding of Windows time?) and w32time service. The later uses the NTP time standard, while older Windows NT 4 time relies on NetBIOS... Hmmm... Maybe that's the reason for a NetBIOS chapter.

I was hoping for something more to deal with Active Directory and modern systems, but learn interoperability to stuff you can only buy at the swap-meet or flea-market.

I didn't cover any other chapters, as they seem to focus on Linux-to-Linux interoperability. Yes, I know, but it is true, Windows coverage in the other sections is less than 5% - 10%.

So bottom line, most of book is Linux with honoroble mention here and there of Windows, the chapters fully focused on Windows interoperability are rehash of SAMBA 2.0 and how to hook into end-of-life Windows NT 4. And you get a whole chapter on NetBIOS? Shall we *cough* look for other books?

This is a very good book on the topic. The author knows what he is talking about, he doesn't skip anything and explains almost everything that should be explained. Although there are some typos and small errors in the book, they are corrected in the errata. Accessible, thorough, well written, well organized, well supported. Definitely worth the money.

Linux in a Windows World by Roderick Smith (O'Reilly) is an excellent aid for anyone trying to work in a heterogeneous environment. It is targeted at administrators who would like to add the flexibility, security and stability of Linux to existing Windows networks. While primarily focused on integrating Linux servers in the enterprise, to add functionality or replace Windows servers, Smith also addresses the addition of Linux workstations and thin clients to a Windows network.

The text begins with a brief discussion of the available options for leveraging Linux in a Windows-dominated environment and when and where such applications are appropriate and effective.
Smith walks you through Samba file and print sharing, authentication with LDAP, Kerberos, NT domains, followed by remote access methods, and a medley of additional server applications for mail, backups and network management. I found the format very accessible; simple and clear enough for those new to Linux, but with sufficient detail to get up-and-running with a basic configuration.

The tone was refreshingly objective. Windows and Linux options were presented in an even-handed way, absent was the Windows bashing that frequently enters Linux literature. I was somewhat disappointed that the book spent so much time on increasingly obsolete NT Domains and so little with Active Directory, but it is understandable as the proprietary nature of Windows necessitates a lag time in the development of interoperable open source technologies.

Overall I found Linux in a Windows World to be an excellent resource for mixed networks. I would consider it a must read for anyone thinking about adding Linux to their Windows network. It covers all the bases, giving the reader an overview of the available options before delving into the most common solutions. I personally found it useful in configuring several servers in my all-Linux network.

This book has well developed details on Linux client configuration for Linux only solutions with some small Windows tidbits thrown in. Given that the scope of this book should be "Linux in a Windows World", I expected more material on connecting Linux to modern Windows infrastructure, especially Active Directory. What coverage that was offered on Windows, e.g. some chapters dedicated to SAMBA, was mostly focused on 10-year older Windows NT era of technologies. I will delve in to some topics to illustrate this.

In Kerberos (Chapter 9), more than 80% of the coverage is about Linux Kerberos solutions and client configuration. There is a measly 2 pages out of 30 on Windows client configuration. In one area (page 216), the author states that Kerberos to Kerberos cross-realm trust to Active Directory would beyond the scope of this book, but I seem to recall that this information took up roughly 5 pages of text in Kerberos, The Definitive Guide by Jason Garman. There is zero coverage on how to configure Linux to use Active Directory KDC, which is something done in about two pages from the same aforementioned book. I'm wondering where's the "Windows World" involved in this as overall coverage is again for Linux to Linux.

For the topic of LDAP (Chapter 8), the coverage is again about configuring Linux clients to use OpenLDAP for authentication. Even within the scope of Linux-to-Linux the information was limited (there's less than adequate coverage about how lack of caching can cause serious problems, or how to use things like SASL for Kerberized authentication of LDAP). There is no mention of using Active Directory LDAP in this chapter for Linux clients. There was some coverage of Windows client configuration through pGina. However, I would prefer to see Windows solutions as this is suppose to be about a "Windows World" book, but get Linux-to-Linux solutions and with pGina, a Windows-to-Linux solution.

On the topic of DNS and DHCP (Chapter 15), only Linux solutions are covered, so there is no information on using Linux with Windows DHCP and DNS backends. In the scope of Linux DHCP, the information is very limited, and there is no information on how to assign IP addresses based on computer names, which is a feature that even Windows DHCP doesn't support. Windows clients spit out names with an appended "0" character to the DHCP server, which could be use for further configuration of services. In the scope of Linux DNS, there is no information about features needed in BIND to make it interoperable with Windows Active Directory, or even Linux Kerberos solutions for that matter. One needs Dynamic DNS, support for SRV records, and in particular for Active Direcotry, one needs support for underscore characters "_", which I heard is a compile time option for BIND9. Additionally, Microsoft has a proprietary WINS record for legacy support that is not supported by BIND. Both Microsoft certification publications and "DNS and BIND" by Albitz and Liu (O'Reilly) offer solution scenarios for this. None of this important essential information was presented, as the predominant focus appears to be Linux-to-Linux.

On the topic of NTP services (Chapter 15), the information is applicable for older Windows 9X/ME/NT systems that used a NetBIOS to locate a non-NTP service from Windows NT. In Windows XP/2003 that is so ancient history and Windows has full support for NTP client and services through the w32tm command. This is completely absent from from the presentation and details on this topic. Also surprising is no mention of why time sychronization is vital to either any Kerberos solutions, including Active Directory; any clients authentication to a Kerberos KDC should sychronize to the clock on that same server, or else users will not be able to log on if their clocks are more than 5 minutes off.

When covering mail technologies (Chapter 13), most of the documentation is an overview of mail technology and a exploration of Linux solutions with sendmail and postfix, which is very good presentation of those technologies. Microsoft Excahnge Server gets an honorable mention of less than 2 pages out of 51 pages. I would have appreciated more scenarios on how to integrate Exchange including for a variety of technologies, and would also have been interested in Windows clients that would hook into Linux solutions. There's no mention that that modern versions of Exchange are dependent on Active Directory -- something of a headache for Linux admins, but important consideration for planning the infrastructure. The Windows interoperability in this area like other chapters was quite lacking.

There are four chapters in Part II "Sharing Files and Printers" dedicated to SAMBA and related technologies like CUPS integration. The coverage here is predominantly focused on 10-year old Windows NT era technologies. One chapter is even dedicated to NetBIOS Network, something even Windows admins wish would have never existed. I don't know but I would hope that interoperability would cover modern technologies that are at least applicable within the last three years (after all, Windows 2003 is about 3 years old now). There only mention of Active Directory, and NONE of the richness found in where Linux boxes can be added to Active Directory domain, authenticate through ActiveDirectory via windbindd (with support for caching), and Windows account sid (security id) to user id (uid) and group id (gid) mapping with LDAP Unix schema support with either SFU 3.5 or AD4Unix in Active Directory (( NOTE: This can be done through SAMBA 3.0.20 and above, but authors haven't a clue )). The interoperability here is for historical network scenarios that is well covered in existing published (and online free) books. I would expect that a book written in this time would offer appicable Windows interoperability.

In the scope of printing through CUPS (Chapter 4), there's no mention that I can tell of Windows support for IPP, the same technology that CUPS uses. Linux clients can connect directly to Windows IPP printer without even touching SAMBA (just have to use port 80 in the URL as Microsoft implemented their solution before the RFC was finalized). Some redeeming qualities of this section was how to do free PDF-Gernation printer share on Linux, and also how to implement Microsoft RPC facility (through SAMBA) for auto-insstallation of printer drivers on Windows clients.

Overall, given that the focus is predominantly Linux for Linux solutions, this book maybe should be called "Linux in a Linux World". If you need an overview of Linux technologies and solutions, then this is a decent book, but if you want modern real-world details on interoperability to Windows, then there are far better books in this space.