Re: Nextgres? - Mailing list pgsql-hackers
From | Justin Clift |
---|---|
Subject | Re: Nextgres? |
Date | |
Msg-id | 3D1A8A0D.467D324A@postgresql.org Whole thread Raw |
In response to | Nextgres? (Justin Clift <justin@postgresql.org>) |
List | pgsql-hackers |
Hi everyone, This is Jonah's explanation of what Nextgres is, as his response didn't make it to the list (some kind of software or network problem). :-) Regards and best wishes, Justin Clift -------- Original Message -------- Subject: Re: [HACKERS] Nextgres? Date: Wed, 26 Jun 2002 12:18:57 -0400 (EDT) From: <jharris@nightstarcorporation.com> Reply-To: <jharris@nightstarcorporation.com> To: "Justin Clift" <justin@postgresql.org>,<pgsql-hackers@postgresql.org> Justin, NEXTGRES is not an extension of PGSQL and is not based on PGSQL code. However, many of the methods used in PGSQL, Oracle, and similar systems have been merged into it. The MVCC engine is similar to PGSQL's and Oracle's in its statement/transaction-level read consistency and other items... whereas have implemented a LRU engine more closely resembling Oracle's. We are taking the best of all systems and incorporating it into one system. As for the SQL, you are correct. There are a few misspellings on the page, and for that I'm sorry. It was last updated *very* early in the morning. We have added support for PostgreSQL's SQL grammar (from 7.1.2). We have a SQL Dialect Interface (SDI) that allows the DBA or user at run-time to execute a SQL block using SQL specific to different database systems. What we have done is take the SQL parser and abstracted it so that during query-rewrite we convert it into our native syntax, which more closely resembles PostgreSQL's with function extensions from Oracle (i.e. NVL, decode, etc.) System catalogs and tables are also aliased so that you could execute an Oracle "SELECT object_name FROM all_objects WHERE rownum < 50" and it is converted to read the similar attribute from our aliased "all_objects" relation. I would be glad to support the open-source effort of PostgreSQL. It has performed much better than any other open-source database Ive ever used. I have personally thanked M. Stonebraker for the original POSTGRES and would like to thank you all for the hard work put into PGSQL over time. I know you guys have representatives for the system, but if you ever need another person to join you at a conference or similar event, just let me know. I look forward to using PGSQL in the future and am sorry that Im too busy to assist you on the backend. -Jonah -- Original Message -- From: Justin Clift <justin@postgresql.org> To: "Jonah H. Harris" <jharris@nightstarcorporation.com> Send: 12:07 AM Subject: [HACKERS] Nextgres? Hi Jonah, Was just looking around your company website, and it mentions a product called "Nextgres" which looks interesting : http://www.nightstarcorporation.com/?op=products How do you guys implement the PostgreSQL SQL parser as well as the Interbase and Oracle parsers? Is it like an adaption of PostgreSQL with addons or something? Also it mentions its compatible with PostgreSQL 7.2.2, so I'm wondering if that's a typo or something. :-) Regards and best wishes, Justin Clift "Jonah H. Harris" wrote: > > Comparing PGSQL to MySQL is like apples to oranges. I don't see why one > would want to take a great project and ORDBMS such as PGSQL and make a > desktop version of it. When a desktop version is completely opposite of > what PGSQL is, a commercial-grade RDBMS. Sure it lacks some of the areas > when compared to Oracle and SQL Server... but I don't see how the PGSQL team > is going to get as much money as Oracle/Microsoft to develop, perform R&D, > and compete against commercial rivals. Yet, I have never seen an > open-source database system as good as PGSQL, especially being as it is > developed on a volunteer basis. > > As far as MySQL goes, they can have their easy-to-install and manage > "features". I was on the MySQL-dev team for three months trying to convince > Monty, Sasha, and others that MySQL needed features found in commercial > systems (triggers, stored procs, transactions, constraints, etc.) They > explicitly and rudely told me that MySQL wasn't developed to perform in > these areas and to go elsewhere. Ever since then, I've been using PGSQL in > a production basis. The argument for easy-to-install systems is common with > many MySQL users, and those who don't understand how databases work. Sure > it would be nice to have the system do complete self-tuning but in reality, > the DBA should know how to make the database perform better under different > situations. And, as for ease-of-install, I can download the PGSQL package > for my OpenBSD boxes and it works perfectly, same on CYGWIN. If I want to > tune it, I can. > > The objective of a good RDBMS is to allow fast access to data while also > maintaining data integrity (ACID properties). I personally think that > dumbing-down database systems only causes more problems. Look at Microsoft > and NT/2K/XP. Now there are MCSEs all over the place acting like they are > network admins because they can point-and-click to start a IIS service. > Oooh, ahh. I would rather be on UNIX where I need to know exactly what's > going on. And, UNIX users don't just jump up and blame the software when > something goes wrong... as often happens with Windows and Access. The same > follows with many MySQL users I've encountered. They don't have to do > anything with the system, but consider themselves experts. With all my > Oracle, SQL Server, and PostgreSQL boxes, I personally tune them to do what > tasks are designated for them. I think PGSQL, as the project goes, is just > fine as it is. A little commercial support and marketing could greatly > assist in furthering the usage of PGSQL, true. If the group agrees that > this would be a good idea, then I would be willing to do this. I also think > it would be a good idea to get a PostgreSQL foundation or similar non-profit > that could accept donations, etc. to further development. Don't dumb down > the system and create a limited version just for people that want an > open-source Access... they can use MySQL for that. Just my rant. > > Cordially, > > Jonah H. Harris, Chairman/CEO > NightStar Corporation > "One company, one world, one BIG difference!" ___________________________________ NightStar Corporate Web-mailer. Open Source for Open Minds.
pgsql-hackers by date: