I've been trying to get a copy of this for a week but it's not in stock either at Borders or Powell's. Based on aleonard's review and other comments, I'm afraid it's just another bad writing job by Wendy Goldman Rohm, whose book The Microsoft File was probably the worst-written I've managed to stagger through in the last couple of years. The reason I staggered through it is that she is not enveloped in Microsoft Fog, so there were some pretty good stories and points of view on Microsoft's business practices. But the writing style? Urgh...
Transactions with rollback (we mainly do SELECTs, and because we don't do transactions, we can be much quicker on everything else). We will support some kind of atomic operations on multiple tables, though. Currently atomic operations can be done with LOCK TABLES/UNLOCK TABLES but we will make this more automatic in the future.
On the other hand, at the O'Reilly open source conference David Axmark told me they eventually plan to support the complete ANSI92 SQL spec. So in this case 'not now' may not be 'never.'
Beating on a drum I have worn out doing so, not everyone understands what database people mean when we say transactions. PHBs and product specifiers, particularly. It really just means support of the SQL standard COMMIT/ROLLBACK feature. Operationally, this means a lot of heavy hauling for the database; just ask an Oracle engineer about how they do it (as I did once).
If you don't really really need "transactions," it's worth thinking carefully about how to turn off the feature in the DBMS you're using (some do allow that), or use one like MySQL that doesn't.
MySQL's developers made a deliberate choice here; it optimizes the program for some applications and rules it out for others.
I'm the co-host of the News conference on the Well, where Jim's mass-scribble was discussed and John Schwartz picked up on it and wrote the WPost article. I've known Jim since he started on the Well a decade ago, and think it's perfectly fine for him to do this, and said so at the time. At my request and others, though he was under no obligation to do so, he briefly mentioned his viewpoint on this. Frankly, it's his own damn business as it should be for anyone on the Well. That's the way it has to be if you really "own your own words."
bandy, booter, blair and now jimrutt are not the only ones to do this over the years, and the reasons and context have been discussed extensively. My guess is that many Well members are uneasy about cases like this but support the right of anyone to do so. On the other side, a few have been vehement about the damage to the flow of discussion (and by inference, to the sense of community) that wholesale removal of postings causes, particularly in topics where the person has been quite active. But my feeling has always been, if you can scribble one posting for whatever reason (and nobody I know of questions that feature), then why shouldn't you be able to do that to all of your postings if you so desire?
jef deserves credit for providing a tool that both makes this more convenient to the user and diminishes the impact on other users.
As for jimrutt, alert Slashdot readers know that I have been quite critical of NSI both before and since he came on board. We've tangled over a number of issues over the years on the Well, but once you look at how he thinks and acts and presents his views (modulo the obvious shit-disturbing comments!), he has genuine integrity. I was not happy with the tripartite agreements announced last week between NSI, ICANN and the Department of Commerce: I think Commerce gave away the store to NSI. But I can't fault Jim for reaching an agreement that is as favorable as he could get for his company. When he first took the job I urged him to focus on dropping the whois database intellectual property claims, and instead get with fixing NSI's obviously broken customer service. To a large degree he is doing that; I'm still not happy with NSI but we have what we have. If the ongoing NSI/ICANN/open registry system can sustain Jon Postel's vision of the net as a resource freely and fairly available to all people, then Jim's professional activities will be vindicated.
But regardless of that, he had the perfect right to scribble his stuff on the Well. Or not.
If y'all are interested in checking out the rather unique place known as the Well, please come check it out. There are both graphical (Engaged) and command-line (picospan) user interfaces, and a lot of ferment in our conferences, and though it's not quite the hotbed of whatever that we used to be, it's still something of a lab of social online interaction.
I'd rather read almost anything than solipsistic scribbles like this -- even how Jon Katz has conquered Linux. I agree with the rest of the peanut gallery: bad call, Hemos. Better to get the notes, humble as they might be, of any other random conference attendee than to plow through this drivel; at least that way interesting issues could be discussed without having to sidestep the ego.
This piece was weak on social insights and nil on technical insights. In addition, Glass has an "illusive" grasp of spelling.
Finally, to cap it all he proposes bringing in the Guardian Angels or something to police the net. Erm, Brett, they already tried. Even as a ha-ha joke this is a bad thing to bring back up.
I've been waiting for someone to come along and say this. I'm still using Word for Windows 2.0 for precisely these reasons. It just works. Actually I still like the DOS Word 4.0 and keep a copy of it, it's handy at times when I need something quick.
If Word's "progress" in the 1990s isn't the classic example of Fred Brooks' Second System Effect, I don't know what is.
The Web is getting increasingly encrusted with junky pseudo-apps that don't really do very much useful. Certain well-known search sites have evolved in that direction also.
I find that Google is just great at nailing down quick references to things. I type in "Georgia counties" or something for some work project and find pretty much what I need immediately.
I would pay folding money for them to keep it this simple and good.
I was at the O'Reilly conference presentation by Monty Widenius and David Axmark last week, and had a long conversation with David afterwards.
GPLing MySQL is a gesture of good faith in the direction of free/open source software while not breaking the model that has brought MySQL undeniable success. They estimate there are somewhere between 200,000 and 1 million copies of the program in use. Given that it's only been available for about four years, that's pretty impressive. It just didn't seem possible given how saturated the database market is.
MySQL is engineered for speed and reliability. The exclusion of transactions, foreign keys, subselects and other features does not simply "make it faster," according to what David told me. The point was to give Monty more time to optimize for speed, robustness and standardization. MySQL has a far more complete adoption of SQL92 than mSQL, for example. They are also committed to adding the remaining standard features such as SQL COMMIT/ROLLBACK (i.e. "transactions").
Far more important to me as a longtime DBA is the availability of subselects. I have few if any apps where COMMIT/ROLLBACK is really operationally necessary, and I'm plenty happy not to have the overhead that SQL "transactions" necessitates. But I really do need subselects all the time for convenience and efficiency. [A subselect is a query in the form select column1, column2... columnx from table y where columnx in (select columnx from table z where...)] It has been near the top of the MySQL to-do list for a year; David says Monty has been cleaning up other stuff as a foundation to implement it soon.
I wouldn't use MySQL to run a heavy-duty banking or ecommerce system, but neither would I use Oracle to run a moderately sized billing system, parts inventory or statistical research in most cases. But let's be clear about one thing: MySQL is only a "lightweight" system in comparison to the investment in time, money and hardware necessary to run a truly "enterprise-level" database. In all other respects it is a commendable effort and I am really pleased that Monty, David and the people who have developed and support MySQL have had so much success with it.
I had actually just bought the book yesterday morning, and turned up at the MySQL presentation at the O'Reilly conference in Monterey in time to hear David Axelrod mention it as a bit disappointing. I'd have to agree.
For one thing, fully one-third of the book is basically man page extracts and API lists (67 pages ALONE for JDBC!!!). Secondly, while database creation and functions are competently developed, there is very little at all about practical issues and approaches to real-world problems. Compare this with, say, O'Reilly's own Annoyances volumes or the Nutshell books.
The lack of familiarity of the authors with the history and extent of the SQL standard were readily apparent. If this were admitted, though, they would have had to highlight the fact that mSQL falls far short in this regard, where MySQL does an excellent job. For example, the authors state on p.234:
"The [MySQL] CREATE INDEX statement is provided for compatibility with other implementations of SQL. In older versions of SQL this statement does nothing."
This is either plain wrong or sloppy editing. CREATE INDEX is the correct statement in the ANSI SQL92 standard, and variants such as MySQL's ALTER TABLE ADD INDEX are equivalents.
The book was clearly rushed to press, with many obvious typos and some difficult to understand layout decisions. For example, the important distinction between MySQL and mSQL reference sections is minimized to a simple subhead in the middle of p. 269. These should be separate chapters or at least designated subsections in the page headings. Otherwise it's quite easy to end up in the MySQL SELECT section rather than the mSQL one, for example.
Where the book is strong, in fact, is in the discussion of how to use various other tools such as Perl scripts or PHP to work in conjunction with MySQL or mSQL. The discussion of CGI is quite good, for example. So really, the book should have been called something like Web Publishing With MySQL and mSQL, because that is really the stance it takes. Nothing wrong with that, really, and it will be helpful to me in that sense, but I think we will have to wait for Paul DuBois' upcoming book to get a more definitive treatment of MySQL, which really deserves better than this surprisingly uneven O'Reilly offering.
Finally, a comment on the ever-present problem of "transactions." I had a long discussion with David Axelrod about this after the presentation yesterday. While I agree with their strategy of not including transactions in MySQL (at least not yet), because it allowed them to optimize the system in other ways, there is definitely a perception problem.
First, it is important to understand what "transactions" means. It is nothing more nor less than adoption of the SQL92 standard for COMMIT and ROLLBACK. These provide for multiple-statement units of work in database operations to provide referential and data integrity. For example, you don't want a financial transaction in a double-entry system to succeed only partially.
The problem is that "transactions" as defined by COMMIT/ROLLBACK is theoretically sound but practically "expensive" approach. In many cases I would say that the benefits of the formal approach can be matched with other techniques, thus avoiding the overhead which is considerable -- up to an order of magnitude in performance loss is not unthinkable with COMMIT/ROLLBACK. Even I was a bit skeptical about this until I got some insight from an Oracle engineer about the incredibly complex requirements of implementing it.
My guess is that 90% of those who think they need "transactions" would find they don't, and frankly, you can pay for a lot of programmer talent and/or database iron for the price difference between MySQL and Oracle, while accomplishing the same results.
Aside from the criticly bad spelling and the grandiosity -- Seven hundred thousand vulnerabilities, gaping holes, wounds in the skin of our present and future information infrastructures, our dream for a free nexus of knowledge, a prosperous digital economy, where we learn, work, play and live our lives. -- I mean, puh-LEEZE . . . still, some interesting results.
This was right before the big wave of tooltalk advisories came out so it may be somewhat less now.
What is instructive is that, face it, TCP/IP and all the associated dependencies are way too complex but we can't roll this back so get used to it. No reason to give up on tightening things, of course.
For an interesting view from another level, read Stephen Northcutt's new book Network Intrusion Detection (New Riders Publishing).
I just received my new copy of Internet World with my friend Jim Rutt's face on the cover. The headline reads "The Demise of Dot Com: And the political storm ahead for Network Solutions' James Rutt." It's not online at Internet World yet so I can't give a specific pointer.
Anyway, it's about time the Commerce Department started closing on this issue. NSI's position has a certain amount of appeal from a self-dealing point of view, but is completely contrary to the intent and the blackletter of the 1993 agreement. They are supposed to manage the whois database in public trust, not convert it to private intellectual property for their own convenience and profit.
The analogy with phone numbers is wrong. Like it or not, the phone company does own your number, although there are some gray-area issues there too.
This issue is fundamental to the autonomy of the global Internet from control by NSI or any other entity. ICANN has problems too, but they are separate.
Let me state this very clearly: we don't know what NSI's intentions are, so we have to separate speculation from reality. But the possibility exists that a privatized whois database would be the leading edge to privatizing the Domain Name System as a whole. What would we do in 1999 or 2000 to overcome such a development?
I urged Jim Rutt in private and reiterate in public my plea for NSI to drop this issue and get to the business at hand: improving NSI's service to its customers, which is widely and correctly regarded as being crummy. They have many advantages as a result of being awarded "first mover" position in the market by virtue of their current government contracts. They would do well to defend that advantage through superior service rather than lawsuits, political arm-twisting and worse.
My 8-bit experience was limited (a few times on an Apple II with a bad mailing list program called ApplePost and a few months with a DECmate, which was really just the noble PDP8 at the very end of its long life cycle).
But I still have my very first 16 bit machine, in fact two of them, the original Columbia PC, one of the first two workable IBM PC clones. (The other was Compaq, the guys next door had one right away in 1983 but we didn't like the toy screen.) The Columbia was a superbly engineered machine that unfortunately was attached to a superbly clueless marketing department. They overpriced the box, didn't provide much support, and after a lot of flailing, disappeared in about 1985, only to have the name re-emerge years later as a hard disk/storage integrator in Florida (Columbia was originally in Columbia, MD).
I still have all the DOS versions going back to 1.1. Haven't fired up my old Columbias for a long time, but I bet they would run at least DOS 6.22 without a hitch. Only 640K of memory, though, so I wouldn't be able to run Linux, even though it has a whopping huge 10 megabyte hard drive. There are 2 256-K Tecmar memory boards in that box. You don't even want to know how much we paid for those in 1984.
From the Old Fogy desk... (P400, 256 megs of RAM, 4x6gig WD drives)...
I certainly have my problems with how ICANN is running, but those problems can be fixed.
What is going to be hard to fix is NSI's emerging effort to take the DNS private. And if and when they do, then a lot of those net-industry "critics" of ICANN are going to gladly join NSI in the forthcoming industry cartel.
I wonder whether we'll see more of the course taken by my neighbors at Tripwire here in Portland, as described in this article. They have taken what was becoming an accepted and widely specified standard "free" security program, wrapped it up in a proprietary package and repositioned it for a corporate market supposedly "uncomfortable" taking on such a program. This orphans the many sites using tripwire, and encourages splitting the code trees on other similarly situated programs to avoid this outcome in the future.
I think the open source model is strong enough to withstand this sort of dynamic, but it's something to keep an eye on.
Jim Rutt is a long-time acquaintance of mine from the Well. When he was hired out of Thomson to run NSI recently, I hoped things there would turn a corner.
Instead, it's getting worse. As has been pointed out, NSI is basically announcing its claim of compilation copyright on the whois database. This is the same greedy crap that West Publishing does with court rulings, and that many an online service does with the postings of its members.
This is evil and must be stopped.
Jim Rutt is a smart guy. I advised him to drop the claims on whois and get to the business of fixing their broken customer service system. So far he is heading in directly the opposite direction.
Yesterday ICANN announced that 15 candidate registrars have been approved to add to the five existing ones. There are 37 more awaiting certification. Probably 5 to 10 will survive as more than mere niche players. Let's insist on better customer support, service, reliability and lack of greed, and may the best registrar win.
I have been as vociferous a critic of NSI on the issue of the whois database as anyone I know.
But then something interesting happened. They hired someone I know, Jim Rutt, who was the CTO at Thomson, to run the show at NSI. Jim has been on the Well for the better part of a decade. In the early 1980s he worked at The Source, the primordial ISP. He knows the Internet in a way I'd bet most of the greedheads who currently run NSI can only dream about.
The accusations of ICANN having a "pro-corporate" bias fail on two counts: (1) ICANN has just begun its operations and it's hard to discern any such bias so far, arguments about the composition of its board and working groups aside; (2) corporations are, after all, the dominant driving force of the net these days, unlike a decade ago. Until real evidence of bias can be shown, I am writing off all such complaints as fearmongering by self-interested scribblers.
All of this is also a distraction from an issue that matters a great deal today: NSI's appropriation (daylight robbery) of the whois database. It is quite clear that it was never the Department of Commerce's intention for NSI to claim intellectual-property ownership of that data. They were hired to manage that data. They can't even claim they have added value, since from the user's perspective the whois database looks and acts just like it did before they took over in 1993.
Blaming "the government" for this is a cheap, lazy shot. Did it ever occur to those who fling such blithe accusations that maybe people in the government are trying to do a good, responsible job on behalf of the American taxpayer? No, it's easier to belittle something you don't care about and don't understand, isn't it.
The whole point being that the government as well as ICANN has made very strong comments contradicting NSI's current version of how they get to their claim over the whois database.
And that is where Jim Rutt comes in. I don't think any one individual can completely change a company, and NSI has evolved from a small tech services firm into a classic Beltway bandit, replete with self-aggrandizing management and sloppy and uninterested service (how about that, you government bashers! -- here's your sainted private sector at "work"). But the CEO does make corporate policy happen, and Jim knows how to do that.
I am betting he will do the right thing, which is to renounce the NSI claim on the whois database, which is and should be a public trust, and instead focus on cleaning up NSI's act and offering a better-mousetrap service that will have the world beating a dotcom path to its door. That's the only way to win in a truly competitive market. NSI has got to get off the quasi-monopoly gravy train, everyone knows it. I hope Jim follows that path.
So Esther Dyson's shot across the bow is designed to get NSI's attention. Realistically, that's what this is all about.
Gosh, I never thought it would be possible to push so many self-inflating insults and misunderstandings into a mere five-line posting:
>>Oh yeah..who appointed ESR to be the spokesman or all things Open Source ? Himself ?
Well, yes, as a matter of fact Himself did, and you know what, so could you! The fact that he actually has something to say might be of note, however.
>>And why would anybody trademark a concept as pure as Open Source, unless they were in it for the money.
I guess that means you think Bruce Perens is "only in it for the money." That ought to get a good laugh from him. And an apology from the likes of you.
>>It's okay for Mr. Torvalds to trademark Linux. It ain't ok for ESR or any self-appointed hypocrite to trademark anything attached therewith.
But I guess it's OK for any self-appointed A.C. hypocrite to whine about it.
>>ESR ain't no Linus Torvalds or Alan Cox, so don't glorify him for nothing.
This raises the curious issue of exactly who it is who is "glorifying" ESR "for nothing." Actually ESR is rather like Linus Torvalds and Alan Cox in that he has written some good code and managed some good open source projects.
This is not to say I always agree with ESR. I don't, but from what I can tell, having met the man as well as read quite a bit of his writing, "what you see is what you get." No hypocrisy, no double standards, no hidden agendas.
I normally wouldn't engage in responding to this kind of mindless attack, but it bears repeating from time to time that NOT hitting the Submit button is often a good idea.
I think that's incorrect. The Performance Monitor measures at whatever increment you would like.
perfmon is actually an excellent and comprehensive utility that has some very nice features and is actually useful in the real world. I use it for database tuning, for example.
I recommend the book (now some years old) on using perfmon. In retrospect it looks like the last hurrah of the VMS crowd before the Win95 mentality took over NT development.
And the fact that perfmon gets no respect is Yet Another Reason Why
Erm, more like 2200 packages and then some. That was the number at Linux World in March, anyway, where the Debian crew were cat-ing the entire distribution on a screen in their booth:)
I'm getting more than a little annoyed at the slander being casually thrown ebay's way. I know Mike Wilson, their VP of engineering. He is a solid guy and really knows his stuff. Anyone in the database business knows that things go wrong. ebay scaled up from nothing to huge in record time. Is that for nought because they have had problems recently? They moved the bar up a whole order of magnitude on what was considered possible for realtime e-commerce on the net.
So lay the fuck off Mike and his engineering operation until you understand the actual details of what was happening there. People seem to believe that "state of the art" comes about through random acts of kindness or something. No. It comes from learning from mistakes and accidents, and the bigger the flaw, the bigger the step forward that those on the leading edge make to put those things behind.
I agree that Microsoft's leap to take advantage of this in marketing and PR was uncalled for. On the other hand, McNealy's crew is famous for that kind of stuff too.
But the spinmeisters at Sun have nothing on Redmond, and so I still have to say
Slashdot Mirror
I've been trying to get a copy of this for a week but it's not in stock either at Borders or Powell's. Based on aleonard's review and other comments, I'm afraid it's just another bad writing job by Wendy Goldman Rohm, whose book The Microsoft File was probably the worst-written I've managed to stagger through in the last couple of years. The reason I staggered through it is that she is not enveloped in Microsoft Fog, so there were some pretty good stories and points of view on Microsoft's business practices. But the writing style? Urgh...
--------
Well, MySQL never said they would "never" do transaction support. In section F3 of the current documentation, "Some things we don't have any plans to do":
Transactions with rollback (we mainly do SELECTs, and because we don't do transactions, we can be much quicker on everything else). We will support some kind of atomic operations on multiple tables, though. Currently atomic operations can be done with LOCK TABLES/UNLOCK TABLES but we will make this more automatic in the future.
On the other hand, at the O'Reilly open source conference David Axmark told me they eventually plan to support the complete ANSI92 SQL spec. So in this case 'not now' may not be 'never.'
Beating on a drum I have worn out doing so, not everyone understands what database people mean when we say transactions. PHBs and product specifiers, particularly. It really just means support of the SQL standard COMMIT/ROLLBACK feature. Operationally, this means a lot of heavy hauling for the database; just ask an Oracle engineer about how they do it (as I did once).
If you don't really really need "transactions," it's worth thinking carefully about how to turn off the feature in the DBMS you're using (some do allow that), or use one like MySQL that doesn't.
MySQL's developers made a deliberate choice here; it optimizes the program for some applications and rules it out for others.
--------
I'm the co-host of the News conference on the Well, where Jim's mass-scribble was discussed and John Schwartz picked up on it and wrote the WPost article. I've known Jim since he started on the Well a decade ago, and think it's perfectly fine for him to do this, and said so at the time. At my request and others, though he was under no obligation to do so, he briefly mentioned his viewpoint on this. Frankly, it's his own damn business as it should be for anyone on the Well. That's the way it has to be if you really "own your own words."
bandy, booter, blair and now jimrutt are not the only ones to do this over the years, and the reasons and context have been discussed extensively. My guess is that many Well members are uneasy about cases like this but support the right of anyone to do so. On the other side, a few have been vehement about the damage to the flow of discussion (and by inference, to the sense of community) that wholesale removal of postings causes, particularly in topics where the person has been quite active. But my feeling has always been, if you can scribble one posting for whatever reason (and nobody I know of questions that feature), then why shouldn't you be able to do that to all of your postings if you so desire?
jef deserves credit for providing a tool that both makes this more convenient to the user and diminishes the impact on other users.
As for jimrutt, alert Slashdot readers know that I have been quite critical of NSI both before and since he came on board. We've tangled over a number of issues over the years on the Well, but once you look at how he thinks and acts and presents his views (modulo the obvious shit-disturbing comments!), he has genuine integrity. I was not happy with the tripartite agreements announced last week between NSI, ICANN and the Department of Commerce: I think Commerce gave away the store to NSI. But I can't fault Jim for reaching an agreement that is as favorable as he could get for his company. When he first took the job I urged him to focus on dropping the whois database intellectual property claims, and instead get with fixing NSI's obviously broken customer service. To a large degree he is doing that; I'm still not happy with NSI but we have what we have. If the ongoing NSI/ICANN/open registry system can sustain Jon Postel's vision of the net as a resource freely and fairly available to all people, then Jim's professional activities will be vindicated.
But regardless of that, he had the perfect right to scribble his stuff on the Well. Or not.
If y'all are interested in checking out the rather unique place known as the Well, please come check it out. There are both graphical (Engaged) and command-line (picospan) user interfaces, and a lot of ferment in our conferences, and though it's not quite the hotbed of whatever that we used to be, it's still something of a lab of social online interaction.
--------
I'd rather read almost anything than solipsistic scribbles like this -- even how Jon Katz has conquered Linux. I agree with the rest of the peanut gallery: bad call, Hemos. Better to get the notes, humble as they might be, of any other random conference attendee than to plow through this drivel; at least that way interesting issues could be discussed without having to sidestep the ego.
This piece was weak on social insights and nil on technical insights. In addition, Glass has an "illusive" grasp of spelling.
Finally, to cap it all he proposes bringing in the Guardian Angels or something to police the net. Erm, Brett, they already tried. Even as a ha-ha joke this is a bad thing to bring back up.
--------
I've been waiting for someone to come along and say this. I'm still using Word for Windows 2.0 for precisely these reasons. It just works. Actually I still like the DOS Word 4.0 and keep a copy of it, it's handy at times when I need something quick.
If Word's "progress" in the 1990s isn't the classic example of Fred Brooks' Second System Effect, I don't know what is.
-------
The Web is getting increasingly encrusted with junky pseudo-apps that don't really do very much useful. Certain well-known search sites have evolved in that direction also.
I find that Google is just great at nailing down quick references to things. I type in "Georgia counties" or something for some work project and find pretty much what I need immediately.
I would pay folding money for them to keep it this simple and good.
--------
Perk is what gets the coffee ready. Perq is access to the coffee.
---------
Cool. That explains it: No wonder the NT ftp works correctly, the way that much other Micros~1 net-related stuff most definitely does not.
-------
I was at the O'Reilly conference presentation by Monty Widenius and David Axmark last week, and had a long conversation with David afterwards.
... columnx from table y where columnx in (select columnx from table z where...)] It has been near the top of the MySQL to-do list for a year; David says Monty has been cleaning up other stuff as a foundation to implement it soon.
GPLing MySQL is a gesture of good faith in the direction of free/open source software while not breaking the model that has brought MySQL undeniable success. They estimate there are somewhere between 200,000 and 1 million copies of the program in use. Given that it's only been available for about four years, that's pretty impressive. It just didn't seem possible given how saturated the database market is.
MySQL is engineered for speed and reliability. The exclusion of transactions, foreign keys, subselects and other features does not simply "make it faster," according to what David told me. The point was to give Monty more time to optimize for speed, robustness and standardization. MySQL has a far more complete adoption of SQL92 than mSQL, for example. They are also committed to adding the remaining standard features such as SQL COMMIT/ROLLBACK (i.e. "transactions").
Far more important to me as a longtime DBA is the availability of subselects. I have few if any apps where COMMIT/ROLLBACK is really operationally necessary, and I'm plenty happy not to have the overhead that SQL "transactions" necessitates. But I really do need subselects all the time for convenience and efficiency. [A subselect is a query in the form select column1, column2
I wouldn't use MySQL to run a heavy-duty banking or ecommerce system, but neither would I use Oracle to run a moderately sized billing system, parts inventory or statistical research in most cases. But let's be clear about one thing: MySQL is only a "lightweight" system in comparison to the investment in time, money and hardware necessary to run a truly "enterprise-level" database. In all other respects it is a commendable effort and I am really pleased that Monty, David and the people who have developed and support MySQL have had so much success with it.
--------
The only thing more boring than someone screeching "First!" is the inevitable whinefest that follows.
--------
I had actually just bought the book yesterday morning, and turned up at the MySQL presentation at the O'Reilly conference in Monterey in time to hear David Axelrod mention it as a bit disappointing. I'd have to agree.
For one thing, fully one-third of the book is basically man page extracts and API lists (67 pages ALONE for JDBC!!!). Secondly, while database creation and functions are competently developed, there is very little at all about practical issues and approaches to real-world problems. Compare this with, say, O'Reilly's own Annoyances volumes or the Nutshell books.
The lack of familiarity of the authors with the history and extent of the SQL standard were readily apparent. If this were admitted, though, they would have had to highlight the fact that mSQL falls far short in this regard, where MySQL does an excellent job. For example, the authors state on p.234:
"The [MySQL] CREATE INDEX statement is provided for compatibility with other implementations of SQL. In older versions of SQL this statement does nothing."
This is either plain wrong or sloppy editing. CREATE INDEX is the correct statement in the ANSI SQL92 standard, and variants such as MySQL's ALTER TABLE ADD INDEX are equivalents.
The book was clearly rushed to press, with many obvious typos and some difficult to understand layout decisions. For example, the important distinction between MySQL and mSQL reference sections is minimized to a simple subhead in the middle of p. 269. These should be separate chapters or at least designated subsections in the page headings. Otherwise it's quite easy to end up in the MySQL SELECT section rather than the mSQL one, for example.
Where the book is strong, in fact, is in the discussion of how to use various other tools such as Perl scripts or PHP to work in conjunction with MySQL or mSQL. The discussion of CGI is quite good, for example. So really, the book should have been called something like Web Publishing With MySQL and mSQL, because that is really the stance it takes. Nothing wrong with that, really, and it will be helpful to me in that sense, but I think we will have to wait for Paul DuBois' upcoming book to get a more definitive treatment of MySQL, which really deserves better than this surprisingly uneven O'Reilly offering.
Finally, a comment on the ever-present problem of "transactions." I had a long discussion with David Axelrod about this after the presentation yesterday. While I agree with their strategy of not including transactions in MySQL (at least not yet), because it allowed them to optimize the system in other ways, there is definitely a perception problem.
First, it is important to understand what "transactions" means. It is nothing more nor less than adoption of the SQL92 standard for COMMIT and ROLLBACK. These provide for multiple-statement units of work in database operations to provide referential and data integrity. For example, you don't want a financial transaction in a double-entry system to succeed only partially.
The problem is that "transactions" as defined by COMMIT/ROLLBACK is theoretically sound but practically "expensive" approach. In many cases I would say that the benefits of the formal approach can be matched with other techniques, thus avoiding the overhead which is considerable -- up to an order of magnitude in performance loss is not unthinkable with COMMIT/ROLLBACK. Even I was a bit skeptical about this until I got some insight from an Oracle engineer about the incredibly complex requirements of implementing it.
My guess is that 90% of those who think they need "transactions" would find they don't, and frankly, you can pay for a lot of programmer talent and/or database iron for the price difference between MySQL and Oracle, while accomplishing the same results.
-------
I saw the same thing in Las Vegas last December.
Yet another reason why
Aside from the criticly bad spelling and the grandiosity -- Seven hundred thousand vulnerabilities, gaping holes, wounds in the skin of our present and future information infrastructures, our dream for a free nexus of knowledge, a prosperous digital economy, where we learn, work, play and live our lives. -- I mean, puh-LEEZE . . . still, some interesting results.
Notice which vulnerabilities are the most common:
tooltalk 26.1%
bind 18.1%
wu_imapd 15.5% (hel-LOO! anyone hoooome??)
qpopper 12.4%
wwwcount 11.8%
rpc_mountd 10.8%
This was right before the big wave of tooltalk advisories came out so it may be somewhat less now.
What is instructive is that, face it, TCP/IP and all the associated dependencies are way too complex but we can't roll this back so get used to it. No reason to give up on tightening things, of course.
For an interesting view from another level, read Stephen Northcutt's new book Network Intrusion Detection (New Riders Publishing).
-------
I just received my new copy of Internet World with my friend Jim Rutt's face on the cover. The headline reads "The Demise of Dot Com: And the political storm ahead for Network Solutions' James Rutt." It's not online at Internet World yet so I can't give a specific pointer.
Anyway, it's about time the Commerce Department started closing on this issue. NSI's position has a certain amount of appeal from a self-dealing point of view, but is completely contrary to the intent and the blackletter of the 1993 agreement. They are supposed to manage the whois database in public trust, not convert it to private intellectual property for their own convenience and profit.
The analogy with phone numbers is wrong. Like it or not, the phone company does own your number, although there are some gray-area issues there too.
This issue is fundamental to the autonomy of the global Internet from control by NSI or any other entity. ICANN has problems too, but they are separate.
Let me state this very clearly: we don't know what NSI's intentions are, so we have to separate speculation from reality. But the possibility exists that a privatized whois database would be the leading edge to privatizing the Domain Name System as a whole. What would we do in 1999 or 2000 to overcome such a development?
I urged Jim Rutt in private and reiterate in public my plea for NSI to drop this issue and get to the business at hand: improving NSI's service to its customers, which is widely and correctly regarded as being crummy. They have many advantages as a result of being awarded "first mover" position in the market by virtue of their current government contracts. They would do well to defend that advantage through superior service rather than lawsuits, political arm-twisting and worse.
-------
Yes, that would be Laura LaGassa, and it is probable that if you had to choose, she's the coder in the family :)
-------
www.hyperreal.org
-------
My 8-bit experience was limited (a few times on an Apple II with a bad mailing list program called ApplePost and a few months with a DECmate, which was really just the noble PDP8 at the very end of its long life cycle).
... (P400, 256 megs of RAM, 4x6gig WD drives) ...
But I still have my very first 16 bit machine, in fact two of them, the original Columbia PC, one of the first two workable IBM PC clones. (The other was Compaq, the guys next door had one right away in 1983 but we didn't like the toy screen.) The Columbia was a superbly engineered machine that unfortunately was attached to a superbly clueless marketing department. They overpriced the box, didn't provide much support, and after a lot of flailing, disappeared in about 1985, only to have the name re-emerge years later as a hard disk/storage integrator in Florida (Columbia was originally in Columbia, MD).
I still have all the DOS versions going back to 1.1. Haven't fired up my old Columbias for a long time, but I bet they would run at least DOS 6.22 without a hitch. Only 640K of memory, though, so I wouldn't be able to run Linux, even though it has a whopping huge 10 megabyte hard drive. There are 2 256-K Tecmar memory boards in that box. You don't even want to know how much we paid for those in 1984.
From the Old Fogy desk
--------
I certainly have my problems with how ICANN is running, but those problems can be fixed.
What is going to be hard to fix is NSI's emerging effort to take the DNS private. And if and when they do, then a lot of those net-industry "critics" of ICANN are going to gladly join NSI in the forthcoming industry cartel.
And that is a movie I don't want to see.
-------
I wonder whether we'll see more of the course taken by my neighbors at Tripwire here in Portland, as described in this article. They have taken what was becoming an accepted and widely specified standard "free" security program, wrapped it up in a proprietary package and repositioned it for a corporate market supposedly "uncomfortable" taking on such a program. This orphans the many sites using tripwire, and encourages splitting the code trees on other similarly situated programs to avoid this outcome in the future.
I think the open source model is strong enough to withstand this sort of dynamic, but it's something to keep an eye on.
-------
Jim Rutt is a long-time acquaintance of mine from the Well. When he was hired out of Thomson to run NSI recently, I hoped things there would turn a corner.
Instead, it's getting worse. As has been pointed out, NSI is basically announcing its claim of compilation copyright on the whois database. This is the same greedy crap that West Publishing does with court rulings, and that many an online service does with the postings of its members.
This is evil and must be stopped.
Jim Rutt is a smart guy. I advised him to drop the claims on whois and get to the business of fixing their broken customer service system. So far he is heading in directly the opposite direction.
Yesterday ICANN announced that 15 candidate registrars have been approved to add to the five existing ones. There are 37 more awaiting certification. Probably 5 to 10 will survive as more than mere niche players. Let's insist on better customer support, service, reliability and lack of greed, and may the best registrar win.
It won't be NSI, though.
--------
I have been as vociferous a critic of NSI on the issue of the whois database as anyone I know.
But then something interesting happened. They hired someone I know, Jim Rutt, who was the CTO at Thomson, to run the show at NSI. Jim has been on the Well for the better part of a decade. In the early 1980s he worked at The Source, the primordial ISP. He knows the Internet in a way I'd bet most of the greedheads who currently run NSI can only dream about.
The accusations of ICANN having a "pro-corporate" bias fail on two counts: (1) ICANN has just begun its operations and it's hard to discern any such bias so far, arguments about the composition of its board and working groups aside; (2) corporations are, after all, the dominant driving force of the net these days, unlike a decade ago. Until real evidence of bias can be shown, I am writing off all such complaints as fearmongering by self-interested scribblers.
All of this is also a distraction from an issue that matters a great deal today: NSI's appropriation (daylight robbery) of the whois database. It is quite clear that it was never the Department of Commerce's intention for NSI to claim intellectual-property ownership of that data. They were hired to manage that data. They can't even claim they have added value, since from the user's perspective the whois database looks and acts just like it did before they took over in 1993.
Blaming "the government" for this is a cheap, lazy shot. Did it ever occur to those who fling such blithe accusations that maybe people in the government are trying to do a good, responsible job on behalf of the American taxpayer? No, it's easier to belittle something you don't care about and don't understand, isn't it.
The whole point being that the government as well as ICANN has made very strong comments contradicting NSI's current version of how they get to their claim over the whois database.
And that is where Jim Rutt comes in. I don't think any one individual can completely change a company, and NSI has evolved from a small tech services firm into a classic Beltway bandit, replete with self-aggrandizing management and sloppy and uninterested service (how about that, you government bashers! -- here's your sainted private sector at "work"). But the CEO does make corporate policy happen, and Jim knows how to do that.
I am betting he will do the right thing, which is to renounce the NSI claim on the whois database, which is and should be a public trust, and instead focus on cleaning up NSI's act and offering a better-mousetrap service that will have the world beating a dotcom path to its door. That's the only way to win in a truly competitive market. NSI has got to get off the quasi-monopoly gravy train, everyone knows it. I hope Jim follows that path.
So Esther Dyson's shot across the bow is designed to get NSI's attention. Realistically, that's what this is all about.
--------
Gosh, I never thought it would be possible to push so many self-inflating insults and misunderstandings into a mere five-line posting:
>>Oh yeah..who appointed ESR to be the spokesman or all things Open Source ? Himself ?
Well, yes, as a matter of fact Himself did, and you know what, so could you! The fact that he actually has something to say might be of note, however.
>>And why would anybody trademark a concept as pure as Open Source, unless they were in it for the money.
I guess that means you think Bruce Perens is "only in it for the money." That ought to get a good laugh from him. And an apology from the likes of you.
>>It's okay for Mr. Torvalds to trademark Linux. It ain't ok for ESR or any self-appointed hypocrite to trademark anything attached therewith.
But I guess it's OK for any self-appointed A.C. hypocrite to whine about it.
>>ESR ain't no Linus Torvalds or Alan Cox, so don't glorify him for nothing.
This raises the curious issue of exactly who it is who is "glorifying" ESR "for nothing." Actually ESR is rather like Linus Torvalds and Alan Cox in that he has written some good code and managed some good open source projects.
This is not to say I always agree with ESR. I don't, but from what I can tell, having met the man as well as read quite a bit of his writing, "what you see is what you get." No hypocrisy, no double standards, no hidden agendas.
I normally wouldn't engage in responding to this kind of mindless attack, but it bears repeating from time to time that NOT hitting the Submit button is often a good idea.
--------
I think that's incorrect. The Performance Monitor measures at whatever increment you would like.
perfmon is actually an excellent and comprehensive utility that has some very nice features and is actually useful in the real world. I use it for database tuning, for example.
I recommend the book (now some years old) on using perfmon. In retrospect it looks like the last hurrah of the VMS crowd before the Win95 mentality took over NT development.
And the fact that perfmon gets no respect is Yet Another Reason Why
Erm, more like 2200 packages and then some. That was the number at Linux World in March, anyway, where the Debian crew were cat-ing the entire distribution on a screen in their booth :)
----------
I'm getting more than a little annoyed at the slander being casually thrown ebay's way. I know Mike Wilson, their VP of engineering. He is a solid guy and really knows his stuff. Anyone in the database business knows that things go wrong. ebay scaled up from nothing to huge in record time. Is that for nought because they have had problems recently? They moved the bar up a whole order of magnitude on what was considered possible for realtime e-commerce on the net.
So lay the fuck off Mike and his engineering operation until you understand the actual details of what was happening there. People seem to believe that "state of the art" comes about through random acts of kindness or something. No. It comes from learning from mistakes and accidents, and the bigger the flaw, the bigger the step forward that those on the leading edge make to put those things behind.
I agree that Microsoft's leap to take advantage of this in marketing and PR was uncalled for. On the other hand, McNealy's crew is famous for that kind of stuff too.
But the spinmeisters at Sun have nothing on Redmond, and so I still have to say