Re: Senior Oracle DBA now Learning Informix by Request

Double Echo wrote: > Ben, that's an interesting point you bring up. > > ISAM is the core of MySQL, which seems to be doing quite well in its > space. > > > The beauty of ISAM as a product is that it is a great entry point for > a customer to get into more sophisticated products. Informix > Standard Engine for example, gets a customer excited about a great > product, and allows them to be successful without a lot of > hand-holding. Once they have success with it, they learn there are > "better" offerings such as IDS. This is why I still don't understand > the blindness of IBM to the success of PHP/MySQL, and how they have > missed the nexis Standard Engine has with MySQL. IBM just doesn't > get it. Not every customer needs an IDS shared memory dbspaced > engine. MySQL is proof of that. It can run circles around > MS-SQL-Server, DB2, and I'd even venture to say, Informix, simply > because the install takes less than 2 minutes, the docs are easy to > find, the tools are simple, and the customer can be up and running, > developing, all for the total cost of nothing but the hardware and > infrastructure to run it. But MySQL has limits, and when you start > working with big data, it can get sluggish, slow down, and other > solutions start coming into play. IBM needs to get with the program > and stop trying to shove one-choice solutions out there with no plan > or idea of how today's market is really working. Instead of only > offering IDS, they need to really pump up the volume on SE as a great > place to start. I've been coding a bit for MySQL recently for another project. Why MySQL? Well it's a web site to be hosted cheaply by a third party so the choices are MySQL or MSSQL for more money. That choice is a no-brainer really especially as I want to use open-source development tools. I have to say I like MySQL: it does the job on a small-scale project and the administrative tools are again free and simple to use. PHPMyAdmin is particularly good as the standard graphical tools are buggy. Documentation was easy to find and very good. Informix needs some good simple tools that are at least free with the DBMS. The current tools are in many ways excellent but ISA gets into a lot of complicated stuff - too much for beginners IMHO - and ServerStudio is heavily restricted in functionality unless you pay a non-trivial licence fee. I never needed to set it up as it came pre-installed with Linux and the third party provided everything at their end. As it was a small scale project I didn't worry too much about how the data was stored. I agree IBM could do with a good free offering to try and get into this market and maybe SE is it as it's already been developed. I am sceptical that they could make a big dent in it though as most ISPs are already supporting two DMBSs and probably don't want a third. Also what's the selling point? That its bigger brother IDS is an excellent scalable database? Well it is but that's a fourth DBMS to support and I just don't think the user demand is there. MSSQL is in there because of Visual Studio integration and because many companies policy is to use MS no matter what. I realise that ISPs are not the only game in town but this is an example of what I am familiar with and what I believe to be a big chunk of MySQL's market. I think IBM could only reasonably make SE open source if it was going to generate more revenue than it would lose and wouldn't reveal too much about IDS. I think it would need to be open source, by which I mean GPL, rather than just free to use so that it could be included in the major Linux distributions and get wide exposure. Sun has just changed the Java licence to facilitate precisely this. Ben.

Ben Thompson wrote: > Double Echo wrote: >> Ben, that's an interesting point you bring up. >> >> ISAM is the core of MySQL, which seems to be doing quite well in its >> space. >> >> >> The beauty of ISAM as a product is that it is a great entry point for >> a customer to get into more sophisticated products. Informix >> Standard Engine for example, gets a customer excited about a great >> product, and allows them to be successful without a lot of >> hand-holding. Once they have success with it, they learn there are >> "better" offerings such as IDS. This is why I still don't understand >> the blindness of IBM to the success of PHP/MySQL, and how they have >> missed the nexis Standard Engine has with MySQL. IBM just doesn't >> get it. Not every customer needs an IDS shared memory dbspaced >> engine. MySQL is proof of that. It can run circles around >> MS-SQL-Server, DB2, and I'd even venture to say, Informix, simply >> because the install takes less than 2 minutes, the docs are easy to >> find, the tools are simple, and the customer can be up and running, >> developing, all for the total cost of nothing but the hardware and >> infrastructure to run it. But MySQL has limits, and when you start >> working with big data, it can get sluggish, slow down, and other >> solutions start coming into play. IBM needs to get with the program >> and stop trying to shove one-choice solutions out there with no plan >> or idea of how today's market is really working. Instead of only >> offering IDS, they need to really pump up the volume on SE as a great >> place to start. > > I've been coding a bit for MySQL recently for another project. Why > MySQL? Well it's a web site to be hosted cheaply by a third party so the > choices are MySQL or MSSQL for more money. That choice is a no-brainer > really especially as I want to use open-source development tools. > > I have to say I like MySQL: it does the job on a small-scale project and > the administrative tools are again free and simple to use. PHPMyAdmin is > particularly good as the standard graphical tools are buggy. > Documentation was easy to find and very good. Informix needs some good > simple tools that are at least free with the DBMS. The current tools are > in many ways excellent but ISA gets into a lot of complicated stuff - > too much for beginners IMHO - and ServerStudio is heavily restricted in > functionality unless you pay a non-trivial licence fee. > > I never needed to set it up as it came pre-installed with Linux and the > third party provided everything at their end. As it was a small scale > project I didn't worry too much about how the data was stored. > > I agree IBM could do with a good free offering to try and get into this > market and maybe SE is it as it's already been developed. I am sceptical > that they could make a big dent in it though as most ISPs are already > supporting two DMBSs and probably don't want a third. Also what's the > selling point? That its bigger brother IDS is an excellent scalable > database? Well it is but that's a fourth DBMS to support and I just > don't think the user demand is there. MSSQL is in there because of > Visual Studio integration and because many companies policy is to use MS > no matter what. > > I realise that ISPs are not the only game in town but this is an example > of what I am familiar with and what I believe to be a big chunk of > MySQL's market. > > I think IBM could only reasonably make SE open source if it was going to > generate more revenue than it would lose and wouldn't reveal too much > about IDS. I think it would need to be open source, by which I mean GPL, > rather than just free to use so that it could be included in the major > Linux distributions and get wide exposure. Sun has just changed the Java > licence to facilitate precisely this. > > Ben. The buzz I heard a few years ago about SE was that it was not up for being an open-source product because of some customers using it did not want it to become open source and increase their risk--as in military customers. But I have no idea if this is true or not, just buzz I heard. Also, I don't think SE needs to be open-source to compete with MySQL. I just think it needs to be more recognized, and if anything, be free to use in a production environment just as MySQL is. It doesn't have to have source code exposed, it just needs to be a choice of ISAM-based databases out there for the web community. Perhaps there could be the same kind of offering that DB2 has for-free-usage. MySQL is something a lot of developers can get their arms around quickly without a lot of hand-holding, thusly its success. Informix SE fits nicely in this space as well, with the added benefit of a seamless upgrade to IDS when the data starts getting bigger. MySQL doesn't really have that "upgrade" path, it is what it is, and a lot of MySQL shops have to come up with complicated replication and performance schemes to make it work efficiently with larger and larger data sets. Most web sites that start off with MySQL quickly discover that once they get beyond the development phase of the site that MySQL starts to have problems with big data. Informix doesn't have this problem because you can go from SE to IDS seamlessly, with a clear path to better performance. From a marketing standpoint Informix is ideally suited to the web, but without a market presence it is ( as I say too often ) a wasted opportunity that IBM just doesn't get. Happy Holidays! -DE-