Tech Talk: John Goodson on Database Access Technologies

Discussions

News: Tech Talk: John Goodson on Database Access Technologies

  1. In this interview, John Goodson looks at the proposed changes, performance enhancements and the new connection mechanism in JDBC 4.O. He examines the impact SQL:2003 will have on developers, how applications will use the SQL/XML standard, and the advantages of using a disconnected vs. a connected model for database access.

    Watch John Goodson's Interview


    Note: For those behind corporate firewalls, a new HTTP stream has been made available.

    Threaded Messages (82)

  2. Yikes![ Go to top ]

    John makes some statements that worry me:

    1) "So even in today's internet type world, if you have a JSP, or something that appears to be disconnected, what's really happening is on the backend server, your application connects to a data source, it executes a query, it retrieves the resultsets and then it disconnects."

    I hope nobody writes a web application this way. This is just simple connection pooling.

    2) "So, when you issue a request to connect to a data source, or you retrieve results, you're not hitting the network, which is where your performance penalties are incurred."

    If you want caching, you perform caching. Any O/R engine will do this for you automatically. Additionally, O/R engines detect individual writes and batch them up into single updates with optimistic locking - all automatically.

    3) "Well, again, in JDBC as an example, you're in what's called auto-commit
    mode."

    Now this is just unbelievable. It's nearly impossible to develop an application correctly with the JDBC driver in auto-commit mode. You'd essentially have a very slow, non-transactional system.

    4) "And, in fact, one of our recommendations is to not always use PreparedStatements."

    The problem with this is that the formats for objects like Dates and Timestamps are database vendor specific. This is a flaw in the JDBC interface. If they actually updated the interface to Statement to fix this, you'd see more people using it and avoiding the network call.

    5) "So the biggest thing we would recommend is if you know you're only going to need four columns in a table, put those four columns in the select list and don't limit it at get at a time."

    Wow... I mean wow... Who are the people who write these kinds of applications? With employees like that, who needs enemies?

    It's a shame that the interview was littered with so much useless junior material. There is a little bit of useful information in there (new Connection API, support for SQL 2003, more support for vendor-specific features...), but you have to dig to find it.

    God bless,
    -Toby Reyelts
  3. junior material?[ Go to top ]

    "Desktop theorist to dispense wisdom"

    Ha ha, that was funny! One well-meaning impractical O/R proponent (usch..) tries to teach Data Direkt, yes Data Direkt! The ûber-gurus that "forever" have been making jdbc and odbc drivers and beaten everybody in that area for the last 10 years!

    What will be the next I wonder?

    Regards
    Rolf Tollerud
  4. Fluff interview[ Go to top ]

    tries to teach Data Direkt


    I'm not trying to teach anybody anything. If you can read english, you can see that my post is complaining that the interview was mostly fluff.

    Hummm... Rolf Tollerud sounds like a foreign name, and it appears that you have problems spelling english... Do you need me to translate for you? If you can tell me your native language, I can run the post through Google for you. I'm happy to do anything I can to help out a TSS troll^H^H^H^H^Hfriend :)

    God bless,
    -Toby Reyelts
  5. this is going to be fun[ Go to top ]

    Yes, please do! Especially I like to know more about this exiting O/R mapping technology. :)

    Why shall I use it?
    Grateful for giving me your time.

    Regards
    Rolf Tollerud
    (Novice)
  6. this is going to be fun[ Go to top ]

    Hi Rolf,

    O/R (if used well) brigns you:

    - more performance as OR is very well suited for caching,
    - less code as interaction with database is done by O/R tool,
    - more maintainable, cleaner and more flexible code,
    - your application is not db vendor independent but certanly more portable to other dbs.

    Of couruse, O/R is harder to understand and implement well and is not for junior developers like you, ;) (sorry I just had to say that :))
  7. this is going to be fun[ Go to top ]

    Mileta: more performance as OR is very well suited for caching

    Caching should be judiously thought out and planned already during the coding process..

    integrated with performance, performance, performance all time in mind. It is nothing that can be slapped on as an afterthought.

    Mileta: less code as interaction with database is done by O/R tool

    Is that the reason why the TSS Petshop had 4 times as many LOC as the the .NET Petshop?

    Mileta: more maintainable, cleaner and more flexible code

    You mean the 13 files that is autogenerated for every table?

    Mileta: your application is not db vendor independent but certanly more portable to other dbs

    That can be acheived with alternative methods.

    Just a last question: What O/R tool is the guys at SAP using?

    Regards
    Rolf Tollerud
    (curious)
  8. Yikes![ Go to top ]

    The problem with this is that the formats for objects like Dates and Timestamps >are database vendor specific. This is a flaw in the JDBC interface. If they >actually updated the interface to Statement to fix this, you'd see more people >using it and avoiding the network call.

    Why is this a flaw in the JDBC interface? The flaw is in the databases that doesn't follow the SQL standards. With a Statement you send a statement as it is without parameter markers. Do you want the Statement to parse the SQL and transform it to the "correct" syntax?
  9. KISS[ Go to top ]

    Fredrik: "The problem with this is that the formats for objects like Dates and Timestamps >are database vendor specific"

    It is almost always best to just set the Dates and Timestamps in the database as chars in the international standard date notation (ISO 8601).

    YYYY-MM-DD
    YYYY-MM-DD hh:mm

    Avoid all problems..

    Regards
    Rolf Tollerud
  10. O/R Tools[ Go to top ]

    The negative comments about O/R tools is interesting.

    I have used them all including the big heavy CocoBase and TopLink tools.

    I have also used pure object databases (Poet, db4o) and they are dream to use but most clients are relational db shops.

    So then what are we to do? I found Hibernate about 2 years ago and it allows me to stay in the Java language (not have to write loads of sql) just like I did when using pure object databases. We use templates for DAO which does most of the CRUDS for us. We do add some of Hibernates query language when creating additional finder methods.

    So we use Hibernate and Middlegen to create all of the value objects and mapping files. Then we use Intellij/IDEA templates to create DAO classes for the value objects.

    For stored procs we use a template for a stateless session bean and XDoclet tags to synch the Interfaces and DDs. XDoclet does the rest.

    It kicks maximum booty and saves a heck of a lot of dev time.

    The discussion about disconnected vs connected rowsets is irrelevant as the usage of one versus the other depends purely on what you are trying to do and how you would like to use resources.

    Re: autoCommit set to true by default.... One could understand that a developer new to the JDBC world might very well forget to issue the commit if it was not set to auto commit by default. So big deal - set the autoCommit to false and commit/rollback at your leisure gentlemen.

    If anybody wants my Idea templates and an Ant build file send me a note at coadams at verisign dot com

    Later,

    Cory
  11. KISS[ Go to top ]

    Fredrik: "The problem with this is that the formats for objects like Dates and Timestamps >are database vendor specific"

    >
    > It is almost always best to just set the Dates and Timestamps in the database as chars in the international standard date notation (ISO 8601).
    >
    > YYYY-MM-DD
    > YYYY-MM-DD hh:mm
    >
    > Avoid all problems..
    >
    > Regards
    > Rolf Tollerud

    Care to explain how you deal with date calculations and formatting for i18n using this solution?
  12. KISS[ Go to top ]

    yes store everything in a blob.
  13. KISS[ Go to top ]

    Fredrik: "The problem with this is that the formats for objects like Dates and Timestamps >are database vendor specific"

    > >
    > > It is almost always best to just set the Dates and Timestamps in the database as chars in the international standard date notation (ISO 8601).
    > >
    > > YYYY-MM-DD
    > > YYYY-MM-DD hh:mm
    > >
    > > Avoid all problems..
    > >
    > > Regards
    > > Rolf Tollerud

    Try JDBC escape for date, time and timestamp constants : {ts '2004-01-09 20:11:11.123455'}, before to do semething lame.

    >
    > Care to explain how you deal with date calculations and formatting for i18n using this solution?
  14. Re:KISS[ Go to top ]

    It is almost always best to just set the Dates and Timestamps in the database as chars in

    >the international standard date notation (ISO 8601).

    It is not good if you need to do calculation or format data. But the old trick suggests to keep all time related fields as long (seconds).

    Dmitry Namiot
    Coldbeans
  15. How to do date calculations and formatting (on string dates)?

    I leave this great problem to the audience. I am sure they will find a way! If they are in good form possible in 4-5 codepages ;)

    (Anything less makes a J2EE developer embarrassed)

    Regards
    Rolf Tollerud
  16. How to do date calculations and formatting (on string dates)?

    >
    > I leave this great problem to the audience. I am sure they will find a way! If they are in good form possible in 4-5 codepages ;)
    >
    It is trivial, but it is very lame for many reasons, you are talking like VB developer. Have you ever used databse for something more than single user desktop application ? Are you kidding or think people on this forum are stupid ?

    > (Anything less makes a J2EE developer embarrassed)
    >
    > Regards
    > Rolf Tollerud
  17. O/R mapping == J2EE's holy cow[ Go to top ]

    I was away for a while and failed to notice that several of my post was marked as noisy. Well we are touchy to day are we not?

    I retire for to day then, bye
    Rolf Tollerud
  18. O/R mapping == J2EE's holy cow[ Go to top ]

    Wasn't me that marked the post noisy mate but I can see by your title that you have missed the point of O/R tools.

    OK then.

    If you would accept that Object languages give you huge advantages over procedural/non OO languages (Even MS figured it out with the intro of C# and .Net) then it would be logical to assume that you would want to stay in the OO world when dealing with databases - hence O/R mapping tools. It's not a J2EE holy cow at all as the concept has been around for much longer than J2EE. My professors back in the 20th century used to call it the object / relational impedence mismatch which used to equate to me having to write butt loads of sql code and bind that code to my objects. Talk about tedious boring work.....

    With regard to the Kook and dates:

    java.util.Date someDate = DateFormat.getInstance().parse("12/1/1999");

    Ya I've done that crap with C++ and C and have to rely on a myriad of libraries from various packages and vendors that still do not provide all of the Date features that Java has natively in the language.

    Try rolling days, weeks, months, etc forward and backward with other languages in order to determine something like what day of the week, month and year 2/3/2004 + 689 days would yield with other languages (don't forget those leap years because we just had Feb with 29 days.....). Good luck.

    Cory
  19. O/R mapping == J2EE's holy cow[ Go to top ]

    My professors back in the 20th century used to call it the object / relational impedence mismatch which used to equate to me having to write butt loads of sql code and bind that code to my objects. Talk about tedious boring work.....

    >

    I still use that terminology. Of course one has to with those still in that century.

    The principle also applies to the View - which explains the plethora of fameworks and patterns trying to match the web view "paradigum" (loved the commercial) with the OO middle.
  20. life stinks[ Go to top ]

    Cory: "Try rolling days, weeks, months, etc forward and backward"

    1) As you showed yourself, converting is 1 (one) row of code. After that you use all the methods in the Calendar package to your hearts content. BTW, the Java Calendar class library very much inferior to the C# library!

    2) I think the reason that O/R tools are not popular in the .NET camp (despite of more than 30 products) has something to do with rich clients. The more functional client, the more problems.

    3) In SOA, am I supposed to a. First get the data from the database, b. Then convert to Objects, c. Then convert again to the XML?

    Anyhow it is not my day today. Just had to suffer a strong reprimand from Ted for my posting on the .NET :(

    Regards
    Rolf Tollerud
  21. life stinks[ Go to top ]

    1) As you showed yourself, converting is 1 (one) row of code. After that you >use all the methods in the Calendar package to your hearts content. BTW, the >Java Calendar class library very much inferior to the C# library!


    I would expect that C# could make improvements as it had the benefit of hindsight in that it could take from Java and other languages and add improvements. It's a fine language except for the fact that it runs on only one platform which doesn't work for most of my clients.

    >2) I think the reason that O/R tools are not popular in the .NET camp (despite >of more than 30 products) has something to do with rich clients. The more >functional client, the more problems.

    You may write gui tools when using the MS tools approach for gui development and not need O/R tools because you would use their disconnected recordset which is tightly intergrated with MS Visual Studio. The JBuilder folks have been doing it as well for a long time and it's nice and easy to use.

    But if you do develop a robust 3 tier app and try to use an MVC approach the same O/R needs apply and there are some free tools for C# starting to emerge. I could ask a buddy what he found in his research if you are interested.

    My C# experience only consists of tutoring a friend that was taking a C# class and it was similar enough to Java that I could help her.

    3) In SOA, am I supposed to a. First get the data from the database, b. Then convert to Objects, c. Then convert again to the XML?

    I'm not sure what SOA is so I can't offer any insight.

    With regard to J2EE complexity someone made on another post, there are definately aspects that are overly complex but that may come with the nature of the beast. J2EE does not have the luxury of being a 'product' offered by one vendor. Instead it's a spec that has to take multiple vendor/peoples' views into consideration which can lead to bloat.

    The drawback that even Sun admitted was that they thought that tool vendors would have made much of the development simpler sooner. The tools are there now but CMP2.0 is still very complicated (without the use of tools) and there are people like Richard Monson Haefel(sp?) that are pushing for simplification of parts of the spec.

    Java is not a religion for me as I thought the CSE major was going to help me fly jets for the military but Reagen kicked the Eastern bloc's ass shortly before I went looking for flight slots..... So here I sit not behind the stick of an F-16, but instead hacking Java code which is actually fun. I like the language because it is elegant and powerful and let's me solve problems and model the world around me from a business perspective (and runs on the systems that my clients use in the enterprise).

    Party on and tell me what SOA is.

    gutten nacht,

    Cory
  22. Tollerund or TROLLerund ?[ Go to top ]

    Sheesh...
  23. life stinks[ Go to top ]

    1) As you showed yourself, converting is 1 (one) row of code. After that you >use all the methods in the Calendar package to your hearts content. BTW, the >Java Calendar class library very much inferior to the C# library!


    I would expect that C# could make improvements as it had the benefit of hindsight in that it could take from Java and other languages and add improvements. It's a fine language except for the fact that it runs on only one platform which doesn't work for most of my clients.

    >2) I think the reason that O/R tools are not popular in the .NET camp (despite >of more than 30 products) has something to do with rich clients. The more >functional client, the more problems.

    You may write gui tools when using the MS tools approach for gui development and not need O/R tools because you would use their disconnected recordset which is tightly intergrated with MS Visual Studio. The JBuilder folks have been doing it as well for a long time and it's nice and easy to use.

    But if you do develop a robust 3 tier app and try to use an MVC approach the same O/R needs apply and there are some free tools for C# starting to emerge. I could ask a buddy what he found in his research if you are interested.

    My C# experience only consists of tutoring a friend that was taking a C# class and it was similar enough to Java that I could help her.

    3) In SOA, am I supposed to a. First get the data from the database, b. Then convert to Objects, c. Then convert again to the XML?

    I'm not sure what SOA is so I can't offer any insight.

    With regard to J2EE complexity someone made on another post, there are definately aspects that are overly complex but that may come with the nature of the beast. J2EE does not have the luxury of being a 'product' offered by one vendor. Instead it's a spec that has to take multiple vendor/peoples' views into consideration which can lead to bloat.

    The drawback that even Sun admitted was that they thought that tool vendors would have made much of the development simpler sooner. The tools are there now but CMP2.0 is still very complicated (without the use of tools) and there are people like Richard Monson Haefel(sp?) that are pushing for simplification of parts of the spec.

    Java is not a religion for me as I thought the CSE major was going to help me fly jets for the military but Reagen kicked the Eastern bloc's ass shortly before I went looking for flight slots..... So here I sit not behind the stick of an F-16, but instead hacking Java code which is actually fun. I like the language because it is elegant and powerful and let's me solve problems and model the world around me from a business perspective (and runs on the systems that my clients use in the enterprise).

    Party on and tell me what SOA is.

    gutten nacht,

    Cory
  24. KISS[ Go to top ]

    Fredrik: "The problem with this is that the formats for objects like Dates and Timestamps >are database vendor specific"

    >
    > It is almost always best to just set the Dates and Timestamps in the database as chars in the international standard date notation (ISO 8601).
    >
    > YYYY-MM-DD
    > YYYY-MM-DD hh:mm
    >
    > Avoid all problems..
    >
    > Regards
    > Rolf Tollerud

    But why? There is a standard for how date/times/timestamps should be represented in the database. You really gain a lot when using the date datatypes, like the possibility to calculate intervalls and so on. If you use PreparedStatement the conversion between different format will be done automatically (that it, to store and retreive them, not do do calculation and so on).
    Why should I limit myself because some crappy databases doesn't support the standars?
  25. KISS[ Go to top ]

    Why should I limit myself because some crappy databases doesn't support the standars?


    I have never saw so crappy databases,
    probably Rolf doe's not know about JDBC escapes (it is the same in ODBC and OLEDB, .NET expert must know it too). It is not a database problem, just some lame developers makes it more complicated for themself (It is very common for mouse deriven architecture experts)
  26. J2EE developers drive right into the wall in 120 km/hour, refusing to embrace KISS.

    Last Time I checked for two or three years ago Oracle, MySQL and SQL Server all had different so called "escape sequence". Something with the quotation marks or whatever..

    I never had any problems with setting date as a string. The collating sequence will be correct. BETWEEN will work..

    Why do you love complexity so much? Because it makes you feel superior? Ok, I have the perfect gift to all J2EE developers in TSS! Here it is, "from me to you", exactly what you have missed all this time,
    http://www.elsewhere.org/cgi-bin/postmodern

    Use it to give that extra touch to your documentation and reports.

    "Make it as simple as possible, but not simpler" Einstein

    Regards
    Rolf Tollerud
    (BTW database drivers is some of the most difficult you can do, on par with Rocket-Science. Come to think of it Rocket-Science is probably easier ;)
  27. Last Time I checked for two or three years ago Oracle, MySQL and SQL Server ...

    So you just checked it a few year ago and you have never used database ?
    Read this
    http://msdn.microsoft.com/library/default.asp?url=/library/en-us/odbc/htm/odch08pr_17.asp
    I hope you trust Microsoft, do not you ?
  28. think again[ Go to top ]

    Juozas: "Have you ever used databse for something more than single user desktop application"

    Juozas: "you are talking like VB developer"

    Juozas: "you have never used database"

    So you expect a serious answer from me after these remarks?
    Not even a "thank you" for my little gift?
    Marks my posts as noisy too?

    Sorry,
    Rolf Tollerud
  29. think again[ Go to top ]

    Juozas: "Have you ever used databse for something more than single user desktop application"

    >
    > Juozas: "you are talking like VB developer"
    >
    > Juozas: "you have never used database"
    >
    > So you expect a serious answer from me after these remarks?

    Sorry,
    if it is not true and you are JAVA and RDBMS expert, I do not understand this kind of humor like "store dates as strings for portability",
    only lame MS users use this way.

    > Not even a "thank you" for my little gift?
    > Marks my posts as noisy too?
    >
    > Sorry,
    > Rolf Tollerud
  30. KISS[ Go to top ]

    Fredrik: "The problem with this is that the formats for objects like Dates and Timestamps >are database vendor specific"

    >
    > It is almost always best to just set the Dates and Timestamps in the database as chars in the international standard date notation (ISO 8601).
    >
    > YYYY-MM-DD
    > YYYY-MM-DD hh:mm
    >
    > Avoid all problems..
    >
    > Regards
    > Rolf Tollerud

    Rolf my brother did you really mean that you want to store dates as char values? What about the date functions that most dbs offer as far as comparisons and such?

    Maybe it's just me but I also feel much more comfortable using tools and languages that are strongly typed. I spend less time converting things and chasing down NullPointerExceptions and I can look at an entity diagram and or class diagram and understand the domain more clearly when proper typing is used....

    We are trying to develop an app now based on (and I kid you not) a database that was literally converted from a flat file db to a relational db with no constraints added and chars for numeric data and chars for dates and..... IT NEEDS TO BE TAKEN OUT BACK AND SHOT!!!! All right... It's OK - deep breath.

    Cory
  31. Comedy[ Go to top ]

    It is almost always best to just set the Dates and Timestamps in the database as chars in the international standard date notation (ISO 8601).

    >
    >YYYY-MM-DD
    >YYYY-MM-DD hh:mm
    >
    >Avoid all problems..


    LMAO!


    Rolf, you are priceless :)
  32. Comedy[ Go to top ]

    Ah well, and don't forget to store all Strings just as BIGINTs consisting of concatenated ASCII values - this will surely avoid all Character Set encoding problems.
  33. It’s getting personal..[ Go to top ]

    Hi Gavin, how is it going with your silly little toy Hibernate?
    Don't forget to check out my post here,
    http://www.theserverside.net/discussions/thread.aspx?thread_id=24291#112974
    and here,
    http://www.k-1.com/Orwell/index.cgi/work/summaries/animf.html

    Don't miss "Look no further than the JBoss"

    Bye,
    Rolf Tollerud
    (who likes nothing more than to stick pins in overinflated egos and overhyped products from "Open Source")
  34. It’s getting personal..[ Go to top ]

    Learn to store dates before to talk about O/R mapping and databases.
    Lame posts are just not interesting.

    > Hi Gavin, how is it going with your silly little toy Hibernate?
    > Don't forget to check out my post here,
    > http://www.theserverside.net/discussions/thread.aspx?thread_id=24291#112974
    > and here,
    > http://www.k-1.com/Orwell/index.cgi/work/summaries/animf.html
    >
    > Don't miss "Look no further than the JBoss"
    >
    > Bye,
    > Rolf Tollerud
    > (who likes nothing more than to stick pins in overinflated egos and overhyped products from "Open Source")
  35. I wonder...?[ Go to top ]

    Rolf,

    Have you ever used Hibernate? To refer to it a "toy" would indicate otherwise and you comments about storing dates as chars would further cement my theory that you have never actually seen a database. You are here on a J2EE forum discussing topics that you are quite clearly unfamiliar with, spouting class-war polemics (that you have no idea about I might add) like they actually have something to do with the topic of the thread.

    Your constant references to .NET make me chuckle since it is evident you have not implemented a "real world" application using the technology. Having been involved with .NET since beta 1, my company rolled out a few applications in response to client request - now 12/18 months down the line we are porting 90% of them to J2EE - the reason: despite C#'s obvious elegance the underlying architecture is just not suitable to enterprise applications - I point you towards the ridiculous overhead of COM+ marshalling that must be used just to get distributed transactions. The shockingly buggy MSMQ support (we had to write our own bloody wrapper!). the poor performance of the reflection code in the standard libraries (Delphi.NET is 100 times faster) The fact that ASP.NET is a nice implementation, that Web Services are very well done and that the toolset is very well put together can not override the flaws in the components used by serious applications.

    I await your response no doubt littered with references to "coffee-table" books that have no connection with the topic whatsoever.

    Rob Harrop
  36. as simple as that[ Go to top ]

    I am a little tired of being abused, and marked as noisy just because stating the obvious. Therefore I have put up a comprehensive view of my opinion at,
    http://www11.brinkster.com/monoasp/java-net/opinion.stm

    If you read it, and answer without personal attacks, I will answer you.

    Of course, if you just have opinions, like,
    "the underlying architecture is just not suitable to enterprise applications"
    It is just an opinion, which you are entitled to, but not so much to talk about..show some links and/or benchmarks.

    At my page, you will see that I still admit that "The advantage they have left, a larger pool of experienced OO programmers.."

    But don't forget one thing, the MS programmers also have an advantage- they are very good in coding for performance- as Petshop I, II and III had shown.

    Regards
    Rolf Tollerud
    (Internet is, after all fantastic)
  37. as simple as that[ Go to top ]

    Rolf,

    I appreciate your views, but one thing annoys me slightly - your assumption that all Java programmers are more concerned with Entity EJB and ORM and their perception as intellectuals than performance and scalabilty is incorrect. I admit that a lot of developers think this way but many do not. I point you to a particularly excellent work by Rod Johnson http://www.wrox.com/books/0764543857.shtml which addresses many of the issues you discuss. I think the problem arises with .NET in that it has a very unhealthy "one size fits all" approach. J2EE benefits greatly for the variety of solutions to a problem and the excellent understanding with the developer community regarding the correct use of design patterns. O/R mapping tools are a big thing in the J2EE world but they are not the solution to every problem but they are the solution to a wide range of problems. With .NET developers are limited, at present, to the Microsoft standard method for data access.

    As for proof about my opinion of .NET being suited to enterprise architecture or not - I do not require proof of my own experiences - the proof has come from years of experience developing .NET applications and wrestling with enterprise features that are substandard to J2EE. Understand when I say enterprise features I mean enterprise features because currently .NET has a much better client model than J2EE - Windows Forms is a very elegant technology but then so was Borland VCL and we all know the connection there. However from my experiences with .NET 2.0 there have been no visible moves to address the issues related to enterprise features instead Microsoft have chosen to simplify the toolset even further (which they don't really need to because .NET is soo simple). So whilst the J2EE has identified its weakness (complexity and poor rich client model) and is making moves to address this (Rave, Spring and SWT as examples) Microsoft have chosen to improve .NET in an area where it clearly already excels and have ignored an area where it is lacking.
  38. as simple as that[ Go to top ]

    Rob,

    Thank you for no Ad Hominen attacks.

    Rob: "your assumption all Java programmers are more concerned with Entity EJB and ORM and their perception as intellectuals than performance and scalability is incorrect"

    "All" is a strong word, certainly Vic Cekvenich don't belong for instance. On the other hand too many does. I got into this thread because somebody took the "über-computer-scientist" attitude against the C++ programmers at DataDirect.

    Rob: "I point you to a particularly excellent work by Rod Johnson"

    You don't have to. Juergen Hoeller and Rod Johnson are my "idols".

    Rob: "I think the problem arises with .NET in that it has a very unhealthy "one size fits all" approach"

    Anything that you can do in Java, you can do in C#. It is a Java copy..

    Rob: "J2EE benefits greatly for the variety of solutions to a problem"

    Of which 95% have 1-3 developer in the development team.."

    Rob: "they are not the solution to every problem"

    They are useful for low budget contract consulting :)

    Rob: "the excellent understanding with the developer community regarding the correct use of design patterns."

    I perceive that they are always fighting.

    Rob: ".NET developers are limited, at present.."

    There are over 30 O/R products for the .NET already (there are some "computer theoreticians" in the .NET world too :) In addition when Objecspaces from MS is released, you'll see that it is not exaggeration to call Hibernate a toy.

    Rob: ".NET has a much better client model than J2EE - Windows Forms"

    XAML/Avalon is going to set the new industry standard for both Windows and Rich-Clients.

    Rob: "However from my experiences with .NET 2.0 there have been no visible moves to address the issues related to enterprise features"

    So the Indigo and SOA means nothing to you then?

    And Vic wants me banned! I don't know if he is really serious or he says so because he is afraid of "guilty by association". In fact I know that many J2EE guys agree with me but gives no support for the above reason. I know too that the whole J2EE world is in moving in the right direction because the situation is very different (to the better, no EJB) than last year. Anyhow I am setting up my new site as an assurance. If I get banned, I can comment anything I want. At least it is not possible to ban me from reading! (TSS is my main entertainment- better than TV).

    Regards
    Rolf Tollerud
  39. as simple as that[ Go to top ]

    Rolf,

    Just to address a couple of your points. Firstly as for the claim that anything you can do with Java you can do with C# doesn't really hold up too well for as I sit here bashing away on my Powerbook. Yes I have heard of Rotor and Mono but I still cannot build a reliable GUI app for my Mac with C# - I can with Java.

    Rolf: Of which 95% have 1-3 developer in the development team
    95% eh? How do you figure? And in all fairness some projects don't need more than that - I mean however many guys do you need to write for instance Commons Collections or Commons Logging. If you look at some of the other projects such as Spring (13 developers on Sourceforge), Struts (about 10 active), Hibernate (Gavin King and JBoss need I say more)

    Rolf: So the Indigo and SOA means nothing to you then?
    It is my understanding that these are features of Longhorn not of the .NET 2.0 framework due to be released along with Whideby late 2004 - I don't see how the promise of a product in 2006/7 solves problems now. Right now J2EE has superb support for transactions, superb support for messaging and well designed DB abstraction model, you have to admit that the whole IDataReader thing is pretty flawed and quite as a clean as JDBC (apart from the exception handling). Currently .NET offers poor transaction support via COM+ and ServicedComponent (meaning I lose my one chance at inheritance!), shocking support for MSMQ (maybe they will fix this in 2.0). The fact that MS are promising that longhorn will ship with an SOA framework in 2007 does not make .NET a viable solution now. And lest we forget the Hivemind project is now in full swing over at Apache and that will offer a full SOA solution to the Java world. I was a big fan of C# when it first came out - the language was clean, it has a non-Microsoft administered standard but the standard .NET framework implementation for enterprise techs is poor and Microsoft's C# isn't even standards-compliant!

    Rolf: There are over 30 O/R products for the .NET already (there are some "computer theoreticians" in the .NET world too :) In addition when Objecspaces from MS is released, you'll see that it is not exaggeration to call Hibernate a toy.

    I hope you know that most of those O/R frameworks are poor in comparison to Hibernate I know because I had a write a bloody alternative one for a client. As for Objectspaces - again we can't live off the hype let's see what happens when that arrives in the meantime I will continue to solve real problems with J2EE and Hibernate. It's interesting though because I think that a port of Hibernate to .NET would be much more successful than the Java version. I know that a lot of very smart agree with me - in fact my company was given a research grant to investigate dynamic .NET IL generation for purposes similar to Hibernate.

    Rob
  40. JDBC 4.O and SQL 2003 XML[ Go to top ]

    Rob: "Currently .NET offers poor transaction support via COM+ and ServicedComponent (meaning I lose my one chance at inheritance!), shocking support for MSMQ"

    I am not really the right person to defend COM+ and MSMQ. If it ever should become real that MS overtakes j2EE I suppose heavy duty transactions and messaging would be the last to go. The Final Defense, so to speak. Rest assured that MS has allocated a lot of resources on it! I for one, have made up my mind to become an expert on Indigo. (That BTW will be released separately, before Longhorn, as will Objectspaces, not that I ever will use Objectspaces..)

    But it is a little boring with this ever returning discussion of J2EE vs .NET. Somehow no matter where you are start off, you always end up with distributed transactions. The thread got off a wrong foot after "the interview was littered with so much useless junior material", but let us not hijack it no longer.

    I think that JDBC 4.O and SQL 2003 XML is very exiting technologies, just think of a join of incoming XML data from a web service and a relational database! The one-two punch of first executing a SQL query and then XQUERY the result XML!

    Regards
    Rolf Tollerud
  41. Rolf and ad hominem[ Go to top ]

    Rolf said,

    > "Desktop theorist to dispense wisdom"

    > On the other hand too many does. I got into this thread because somebody took the "über-computer-scientist" attitude against the C++ programmers at DataDirect.

    > The thread got off a wrong foot after "the interview was littered with so much useless junior material"

    Rolf,

    I don't normally say this to people, but I think you need to hear this: With all due respect - shut up. When you have something meaningful to say, in place of the infantile ad hominem attacks you've been directing at me, I'll be happy to listen.

    1) You have yet to explain how any of that material was anything but "junior". Frankly, I'm not surprised that you find it to be informative. The interview appears to be targetting people at your low skill level.

    2) If you had bothered to spend even 60 seconds trying to discover what kind of work I've done, you'd have realized that "desktop-theorist" is about the most incorrect expression you could have possibly used to describe me. Desktop theorists don't write software that processes millions of objects in milliseconds.

    3) You keep on referring to DataDirect employees as elite engineers who have the hidden knowledge of C++ behind them. Since I only know Java, how could I possibly understand the delicate inner workings of a C++ program?

    Well guess what Rolf, I've spent many years of my life writing C++ (including on your beloved Microsoft platform), and I continue to do so even now. You can get a taste of the kind of code I've written by looking at Jace - http://sf.net/projects/jace. It's a C++ runtime library that abstracts JNI. Since you seem to believe most open source is crap, let me pre-emptively cut you off by saying that this software is used by SLAC, CERN, BBC, Carnegie Mellon, Northrop Grumman (yes, the people who build systems for NASA), and many others.

    With the way you behave around here Rolf, the only thing you're achieving is a guarantee that you will never find a job working under or with any of the people who frequent this board.

    Is it possible for a troll to change its attitude? Guess we'll find out in Rolf's next TSS installment.

    God bless,
    -Toby Reyelts
  42. DataDirect engineers[ Go to top ]

    I wrote:

    > 3) You keep on referring to DataDirect employees as elite engineers

    To clarify, this was not to imply that DataDirect employees are incompetent or anything. I hold no personal opinion for or against DataDirect and their products. Somehow Rolf has managed to twist this into a "me vs DataDirect" discussion. The only concern I held was that the interview could have been more appropriately targetted towards the TSS readership.

    God bless,
    -Toby Reyelts
  43. Rolf and ad hominem[ Go to top ]

    You say:
    "Well guess what Rolf, I've spent many years of my life writing C++ (including on your beloved Microsoft platform), and I continue to do so even now. You can get a taste of the kind of code I've written by looking at Jace - http://sf.net/projects/jace. It's a C++ runtime library that abstracts JNI. Since you seem to believe most open source is crap, let me pre-emptively cut you off by saying that this software is used by SLAC, CERN, BBC, Carnegie Mellon, Northrop Grumman (yes, the people who build systems for NASA), and many others."

    I Say:
    I think that since you have stated your credentials and provided a link so anyone can have an idea of who you are and what your work is about, in the name of fairness and fair-play, Rolf should kindly state his credentials and a link to see the type of work he works on.
  44. Marx, Engels, Adam Smith and JBoss[ Go to top ]

    Rolf,

    I've seen the line before from the MS camp saying that Open Source is somehow not only unAmerican but anti-competitive and it's funny to consider.

    From an economic standpoint Open Source has done nothing more than made certain types of knowledge and the applications there of, into commodities (application servers, various software tools, OSs, etc)., which has had the affect of applying downward price pressure to the aforementioned items.

    If anything could be considered Orwellian it would be the exertion of anti-competitive practices on the market by dominant companies. I don't really buy into monopoly theory though, as long as the free market can operate, because I believe competition ultimately will have it's way with the dominant companies within any sector during any particular time.

    Open Source has more to do with the spread/sharing of knowledge and if you fear that you are a cover boy for the antagonist within every Orwell novell. Next you will suggest that all open source projects should be burned -> "Fahrenheit 451"

    Live free or die,

    Cory
  45. Marx, Engels, Adam Smith and JBoss[ Go to top ]

    Hi Cory,

    Cory: open source projects should be burned -> "Fahrenheit 451"

    <respect>Ah, to be well read and educated is a scarcity.</respect>

    All isms, fanatics, fundamentalist groups of any kind, religious or semi-religious, are not my cup of tea.

    If the technology courses in the University had a small amount of Human Arts sprinkled around the students, then they would not be so easy prey to all kinds of movements. The best vaccine against regimes as in Ray Bradbury's "Fahrenheit 451", or thought leaders like Richard Stallman.

    Regards
    Rolf Tollerud
  46. I hope the spec was not too long to read.
    Can you say something about JDBC 4.O/SQL 2003 ?


    > Hi Cory,
    >
    > Cory: open source projects should be burned -> "Fahrenheit 451"
    >
    > <respect>Ah, to be well read and educated is a scarcity.</respect>
    >
    > All isms, fanatics, fundamentalist groups of any kind, religious or semi-religious, are not my cup of tea.
    >
    > If the technology courses in the University had a small amount of Human Arts sprinkled around the students, then they would not be so easy prey to all kinds of movements. The best vaccine against regimes as in Ray Bradbury's "Fahrenheit 451", or thought leaders like Richard Stallman.
    >
    > Regards
    > Rolf Tollerud
  47. 1. There is no "XML Data". XML is a file format, it has no data model. Everyone talking to you about an "XML data model" or an "XML query language" tries to sell you a tool. XML is no different than any of the several thousand other text-file formats, it just has more available "standardized" tools (go figure).

    2. The problem is not "connecting various drivers", it is the crappy quality of most JDBC drivers and the weak SQL standard adoption.

    3. Adding physical pointers such as ROWID to a data access API is a nightmare. It was a bad idea exposing this in Oracle in the first place, please stop it.

    4. "Something disconnected" only gives you better performance if you are happy with the (possibly) stale data in your disconnected Rowset/Objects. It has no real impact on performance, it's a question of application design. If you don't need current data, you of course reduce the number of retrieval operations and use an explicitely cached version.

    5. The stuff about auto commit is indeed hilarious. There are two ways to interact with a database: Planned Statements and Ad-Hoc Statements. An application can never have Ad-Hoc Statements, as all are "planned" by the developer. You execute Ad-Hoc Statements if you sit in front of your crappy SQL (they sell those?) client and execute some reporting queries or insert some test data, ad-hoc. Auto commit mode is ONLY for Ad-Hoc Statement execution. There can never be ad-hoc transaction semantics in an application. The problem is again SQL databases and JDBC drivers: Auto-commit is often enabled BY DEFAULT for new database connections, which confuses Java developers regularly.

    I stopped reading after that, as it is clear that these guys sell tools and not knowledge. Please also check the Hibernate JDBC compatibility page (http://www.hibernate.org/Documentation/CompatibilityGuideSupportedDatabases), there are well known issues with the quality of their JDBC drivers. We actually don't bother anymore with them and recommend other vendors.
  48. Christian: "There can never be ad-hoc transaction semantics in an application."

    Oh!. If I have broken any law I am sorry! I routinely let my customer/clients use a query-by-example form in my web apps. Is that wrong? What law have I violated? Does that mean that your clients can not do their own searches?

    Regards
    Rolf Tollerud
    (confused)
  49. How do search parameters influence the transaction semantics? Are you stupid?
  50. Christian: "Are you stupid"

    Yes.

    I forgot to say that when you have the data in a grid at the client, the users can edit the grid and update it with one click.

    Regards
    Rolf Tollerud
  51. Again, how is this related to transaction semantics? Do you know anything about data management? Your posting history shows that you are probably a troll looking for attention. Do I waste my time here?

    TSS: Can we have a moderation system, please?
  52. Christian: "TSS: Can we have a moderation system, please?"

    Well, that was a little over the top was it not?
    You should see what a guy in TSS.NET is coming away with.

    I have a actual case in mind I would like to ask about- if O/R is a viable option, do you want to take the question?

    Regards
    Rolf Tollerud
  53. If your question is at the same level as your previous statements - No. There is plenty of beginner material around.
  54. probably a wise decision[ Go to top ]

    You know what?
    You made a wise decision.

    Only please stop making comments about DataDirect. Just because you have been involved in a Java O/R project, one of 30-40, all equally idiotic does not qualify you to make any comment about the C++ "real programmers".

    Regards
    Rolf Tollerud
  55. probably a wise decision[ Go to top ]

    You are a cute troll and apparently don't need my help to debunk yourself. Good job.
  56. probably a wise decision[ Go to top ]

    Christian, meet Rolf, our resident troll. He can be quite entertaining in very small amounts.
  57. 3. Adding physical pointers such as ROWID to a data access API is a nightmare. It was

    >a bad idea exposing this in Oracle in the first place, please stop it.

    Is not it a part of SQL-3 standard now?

    Dmitry Namiot
    Coldbeans
  58. I think John Goodson’s interview was balanced and helpful. Even though some of the information may seem trivial for more advanced users of JDBC, nonetheless it can save a few perplexing hours for many developers who are relatively new to JDBC. For example, it is easy to be baffled by funny transactional behavior if the default autocommit setting is true. Similarly, the idea of avoiding a prepared statement if the statement is not going to be reused may not immediately dawn on someone who is seeing the performance benefits of properly using the prepared statements in other parts of the code. Since this interview may be seen/read by all kinds of developers, many of whom may not use an OR-Mapper, such mixture of information has wider appeal.

    I don’t know how DataDirect has kept up with the quality of its JDBC drivers, but back in 1997, when we were developing our OR-Mapping technology (JDX), we found its drivers, along with drivers from WebLogic and Sybase, to be the most stable ones. It was a real pain in those days to build and test the mapping layer using a foundation of unstable JDBC drivers. Most of them did not implement the metadata calls completely and properly. There was a time when many of the apparent bugs could be resolved by just changing the JDBC driver. Anyway, I hope DataDirect continues to help in evolving the JDBC and related standards.

    -- Damodar Periwal

    Software Tree, Inc.
    Simplify Data Integration
    http://www.softwaretree.com
  59. Ban Rolf for ever[ Go to top ]

    It's hard to have a discussion w/ Rolf.

    For example... JDBC JCP does not address our needs, those are valid comments, but I won't debate it becuase Rolf just ruins it.

    .V
  60. Ban Rolf for ever[ Go to top ]

    Vic: "It's difficult to discuss with Rolf"

    Do you mean like the time when I was talking about using Access for disconnected data at the client and you 4 or 5 times had to preach how wrong it was to use Access? (pretending I advocated Access at the server trying to make me look like an idiot?).

    This little trick of pretending to misunderstand is frequently used by another favorite of mine in TSS, no name mentioned.

    But come on Vic, tell us what is wrong with the spec. it look like a large step in the right direction! I am 100% sure that you are no O/R addict, don’t try to fool me.

    Regards
    Rolf Tollerud
  61. Ban Rolf for ever[ Go to top ]

    This little trick of pretending to misunderstand is frequently used by another favorite of mine in TSS, no name mentioned.

    >
    > But come on Vic, tell us what is wrong with the spec.

    Have you read the spec ?

    BTW do not forget to read about dates, it will be very usefull for your.
  62. Ban Rolf for ever[ Go to top ]

    Juozas,

    You need to read this one more time :)
    http://www11.brinkster.com/monoasp/java-net/opinion.stm
  63. Ban Rolf for ever[ Go to top ]

    Juozas,

    >
    > You need to read this one more time :)
    > http://www11.brinkster.com/monoasp/java-net/opinion.stm

    I have read this and I found nothing technical in this crapp. It was very boring to read, sorry.
  64. Ban Rolf for ever[ Go to top ]

    I know more than you about MS products and I have used it in practice ( DCOM, ATL, MFC, MS SDK, ODBC OLEDB, ... ). Microsoft is not an enemy for me. I am not afraid if MS will do something better than SUM or IBM in the next decade, I can become any technology expert very fast (MS JAVA clone is not better at this time, I have tried it too) and you are looser at this time not me. Try to be a star on TSS.NET, you look very lame on this forum.

    BTW Are you paid to advocate unreleased MS software like Longhorn or Object Spaces.
  65. Ban Rolf for ever[ Go to top ]

    Juozas,

    You did not read carefully enough, try again,
    http://www11.brinkster.com/monoasp/java-net/opinion.stm

    "Do you know that Ad Hominen attacks are frowned upon? It is not considered
    Good Taste".
  66. Ban Rolf for ever[ Go to top ]

    Rof,

    Read some book about data access and we will talk. This forum was about data access, but become about your problems to understand data access and about
    great future for MS users. You made this forum about your problems not me.


    BTW
    Try to read this carefully enough: http://www11.brinkster.com/monoasp/java-net/opinion.stm
    It shows error message, try to fix it yourself without my help.
  67. Ban Rolf for ever[ Go to top ]

    This thread was calming down until you started your Ad Hominem attacks.

    You see there is a reason for the rules; as soon as any personal attack starts, the thread disintegrate. So do you thing it was a smart thing to do?
    I have nothing else to do tonight :)

    How long do you want to go on?
  68. JDBC 4.0 and JDBC RowSets[ Go to top ]

    JDBC 4.0 (JSR 221)
    http://www.jcp.org/en/jsr/detail?id=221

    JDBC Rowset implementations (JSR-114)
    http://www.jcp.org/en/jsr/detail?id=114

    jdbc-interest mailing list
    http://archives.java.sun.com/jdbc-interest.html
  69. JDBC 4.0 and JDBC RowSets[ Go to top ]

    Sean, Thanks for the links!

    the Spec..
    "allow a tight association of SQL queries with designated object structures describing query parameters. In a similar linkage, this also allows developers to more immediately tie the results of their queries to designated class definitions and to quickly process a query result"

    <joke>They are stealing from us! Police police!</joke>

    This can only be good, throw away these pesky resultSets.. Then it will be easier than ever to convert code back and forth between Java and .NET.

    ADO Datasets for Java? At least some people will leave OR and use datasets excuse me, rowsets, with stored procedure to get performance and simplified code. Let the heavy duty database take the load, these guys at Datadirect knows what they are doing.

    Suddenly there goes up a light! Now I understand some of the early comments from the O/R guys. My though processes is very slow today (I mean even more slow than usual).

    Regards
    Rolf Tollerud
    (Vic, E Tu Brutus?)
  70. I have sent Dion Almaer a letter (mail) asking if he want me to continue to participate in TSS or not. Awaiting his answer I will only remind the forum that the subject of this thread is the interview with John Goodson about JDBC 4.O/SQL 2003 and "O/R mapping or stored procedures together with rowsets, i e, if Microsoft has been right all the time.

    Regards
    Rolf Tollerud
  71. O/R mapping or stored procedures together with rowsets

    Do you think O/R mapping and stored procedures conflict in some way ?
  72. O/R mapping or stored procedures together with rowsets

    > Do you think O/R mapping and stored procedures conflict in some way ?

    I would be interested in what Gavin might have to say but I have found that they may not conflict but they are not necessarily complimentary.

    For example there may already be a database with stored procs on your next project. Perhaps it is deemed that these stored procs are to be used for performance reasons, your manager likes to inflict pain, etc. Usually the stored procs may represent some specific type of create, read, update or delete type of function but typically not all of CRUDS will be defined for any one entity, using the stored procs....

    So you may then consider using O/R mapping tools for mapping objects to databases with CRUDS for each Class (and the DAO pattern perhaps for your CRUDS) and then consider wrapping stored procs with something resembling the command pattern (Command pattern usually is a class that does one thing and usually provides a set/check method and an execute method). For the commands we currently use SessionBean Intellij/IDEA templates which in turn have a bunch of XDoclet tags. You end up writing very little code. As a side note these stored procs usually returned * from a table so we could reuse the value objects that were generated from the schema using our O/R tool (Hibernate + hbm2java, middlegen).

    Cory
  73. O/R mapping or stored procedures together with rowsets

    > > Do you think O/R mapping and stored procedures conflict in some way ?
    >
    > I would be interested in what Gavin might have to say but I have found that they may not conflict but they are not necessarily complimentary.
    >

    They do not conflict, you can use query rewrite
    and views to make DB look as set of tables for client.
    Logical view of data is one of most usefull RDBMS features.
    I have worked with system implemented as stored procedures,
    I am not sure 100% of procedures are "hidden" in this system (I do not need to know all the parts of this system),
    but it was no direct stored procedure calls or phisical table access on the client for typical use cases.
  74. I have sent Dion Almaer a letter (mail) asking if he want me to continue to participate in TSS or not. Awaiting his answer I will only remind the forum that the subject of this thread is the interview with John Goodson about JDBC 4.O/SQL 2003 and "O/R mapping or stored procedures together with rowsets, i e, if Microsoft has been right all the time.

    >
    > Regards
    > Rolf Tollerud
    The fact that Java will have a spec for rowsets doesn't make MS right, we have to wait for the implementation of the spec to catch in the community. If in the end no one uses the rowset stuff, and keep using O/R for most works, then MS will be proven wrong (what would't be a big surprise anyway).

    Again, I direct you to Fowler's book about enterprise architecture patterns, there you will be able to understand the differences between O/R mapping and rowset, there are situations that require one solution and others the require the other. Just as a hint, rowsets are recommended only for relatively "simple" solutions, and O/R for complex ones the involves having a full domain model in the middle tier. Try to apply rowsets there, and you will surely bury your project 6 feet under.

    BTW, Rolf's credentials can be seen here:
    http://www.theserverside.net/news/thread.aspx?thread_id=24149#112016
    "I never was able to lay my hands on a really big and functional Enterprise application before- they all guard their secrets. I hope everybody download it so we can discuss it. " - Rolf, February 25, 2004

    Regards,
    Henrique Steckelberg
  75. Hi Henrique,

    Have you downloaded the Shadowfax application and seen the scope of it? How many of us have been involved in a project of that scale? (I do not mean in a minor position but being involved in the architecture!).

    Especially as I could never give up my freedom- to work as an employee. The biggest project I ever been responsible for, with the help of 20 Indian programmers, was a 300000 LIC customer loyalty application (a product, not in-house), with ATG Dynamo for three year ago.
     
    BTW, a "well-meaning impractical theorist" doesn't always has to be incompetent (although he usually is). The difference is spelled "deployment". Hence, it is possible to be the world top programmer, and still not understand the pitfalls and difficulties in un-glamourous practical reality.

    Regards
    Rolf Tollerud
  76. Hi Henrique,

    >
    > Have you downloaded the Shadowfax application and seen the scope of it? How many of us have been involved in a project of that scale? (I do not mean in a minor position but being involved in the architecture!).
    >

    Well, I don't know how many of us, but surely _you_ haven't. So the next time you try to "teach" us something, think about it. You come here as a preacher of doom for us, but you don't have the experience to judge what's good or bad for your own enterprise development problems, even less for our problems. You don't have experience with the kind of problems we have to deal with daily, and yet have the "superior wisdom" to judge the solutions we choose to apply to solve them.

    >Especially as I could never give up my freedom- to work as an employee. The >biggest project I ever been responsible for, with the help of 20 Indian >programmers, was a 300000 LIC customer loyalty application (a product, not in->house), with ATG Dynamo for three year ago.

    Three years ago? So actually you don't have any kind of large system development experience in .Net either? How is it possible for you to really know if .Net is such a good enterprise platform without this knowledge? Just by reading MS' marketing fluff?

    > BTW, a "well-meaning impractical theorist" doesn't always has to be incompetent (although he usually is). The difference is spelled "deployment". Hence, it is possible to be the world top programmer, and still not understand the pitfalls and difficulties in un-glamourous practical reality.
    >
    > Regards
    > Rolf Tollerud

    You can take this and apply to yourself: you may be well-meaning (but which troll is?), you have no practice (as stated previsously), and you're a MS' theorist also (some of MS' solutions you preach work only in theory, after all they don't exist yet...), isn't it ironic? ;)

    Regards,
    Henrique Steckelberg
  77. "i am better than you"[ Go to top ]

    Come on Henrique, this mud-casting is boring "I, and my father is a better programmer than you and your father" Are you going to harp on this theme forever?

    This miss-conception of yours that "competences==big projects" is amusing :)

    I did swear to not come into this infected threads again. But the casual dismissal of Datadirect was to much. Do you consider writing a driver be a big project? Datadirect started off, under a different name, as a company that ported db-drivers between all kinds off big mainframe and Unix computers. After more than 30 projects they began to make there own "state of the art odbc drivers". They has been working together with Sun and JDBC from the start and all the way up to the forth-coming JDBC 4.O. They are members of JCP.

    Until some years ago there was a meter standard that was kept in a vault in Paris. The Datadirect guys should also be kept there, as "professional developer standard".

    Regards
    Rolf Tollerud
  78. "i am better than you"[ Go to top ]

    Data direct is the great company and nobody said negative things about it.
    We are taliking about data access, do you want to join ?
  79. "i am better than you"[ Go to top ]

    Come on Henrique, this mud-casting is boring "I, and my father is a better programmer than you and your father" Are you going to harp on this theme forever?

    >

    I am not saying that I am better than you or anyone, this is not a contest. I was just showing you that your attitude regarding most things java is innapropriate and based on false concepts and lack of experience.

    > This miss-conception of yours that "competences==big projects" is amusing :)

    Well, if you want to talk about galaxy formation with an astronomer, you should know something about the subject, shouldn't you? ;)

    >
    > I did swear to not come into this infected threads again. But the casual dismissal of Datadirect was to much. Do you consider writing a driver be a big project? Datadirect started off, under a different name, as a company that ported db-drivers between all kinds off big mainframe and Unix computers. After more than 30 projects they began to make there own "state of the art odbc drivers". They has been working together with Sun and JDBC from the start and all the way up to the forth-coming JDBC 4.O. They are members of JCP.
    >

    I take it you are not an db driver expert, but that doesn't stop you from recognizing their achievment. The fact that you hate Java or some specific technologies related to it, which has solved many problems for most of us in practice, shouldn't stop you from respecting others here who have excelled in these and other areas too.

    > Until some years ago there was a meter standard that was kept in a vault in Paris. The Datadirect guys should also be kept there, as "professional developer standard".
    >

    It's a nice thing that they are working with java too.

    Regards,
    Henrique Steckelberg
  80. But Henrique, I do not hate Java. I am abhorred by the culture.

    In fact, if I should choose or name freely (in fantasy) of all kinds of projects, the one I would be most interested in is: GigaSpaces.

    Regards
    Rolf Tollerud
  81. Rolf: > But Henrique, I do not hate Java. I am abhorred by the culture.
      
    It is true we Java developers are all heathens but we throw the best parties.

    Aren't you being a bit anti-Jamentic?

    Cory
  82. anti-Jamentic?[ Go to top ]

    Arrogant, pompous and quasi-scientific attitude +
    rude, vulgar and impolite manners.

    Is NOT a good combination.

    You are a welcome exception though!

    Regards
    Rolf Tollerud
  83. * query[ Go to top ]

    Does this mean that using JDO(Hibernate)is more performance taxing than using simple JDBC SQL query. Retrieving the whole model requires quering * fields of the table
    Faisal