Comments (6)
  • Add a Comment
  • Edit
  • More Actions v
  • Quarantine this Entry

1 shield commented Permalink

Joe<div>&nbsp;</div> I have not been actively looking at EGL so please be gentle!<br /> I think PHP is getting a lot of buzz in the i space simply because its used everywhere else? I have developed lots of code in PHP to meet my website needs and find the transition from C (my primary programming background) to PHP very easy. If you need to get hands on a project having a bigger pool of developers to choose from is the best option in my view. There are literally thousands of PHP programmers out there so the choice starts to get made for you. EGL sounds like a great language, so did Java but look at how that has languished to the depths.... <br /> PHP has many problems, one of the biggest is IBM's integration on the i, this is not a problem with the language simply the i is holding the language at arms length. Having spent months trying to find ways around simple operations on other platforms which simply do not play well on the i implementation shows just how hands off IBM has been with this! The setup of PHP on the i takes a really good understanding of many technologies which is why it is taking off a lot slower than it should!<div>&nbsp;</div> As I have explained I really don't know enough about EGL to make a comment about its benefits over PHP? Does it run on all other platforms, is the code which is generated platform independent? What applications do I have to purchase from IBM to get it up and running effectively (PHP is free!). How much effort is required to turn my base HTTP server into a fully functional EGL capable server and what is the footprint like? (I had to turn off *ADMIN HTTP server because I don't have enough memory installed! It is a pig and getting fatter).<div>&nbsp;</div> No offence meant, I just think PHP is more than just white noise and if IBM really got behind the technology by creating a fully integrated version we might see more noise?<div>&nbsp;</div> Chris...

2 JoePluta commented Permalink

Interesting points, Chris, but I disagree with some of them on a very fundamental level. For example, I don't consider Java to be "languishing"; it's the most-used programming language in the world. It doesn't get quite as much traction on the i as other platforms, but that's because we have RPG and COBOL to write our business logic in. Java for the most part has two purposes in the i world: web applications and deploying Java packages (like JavaMail or Jasper). Given the number of folks using WebSphere and/or Tomcat, there's a lot of Java out there, and given how easy it is to invoke Java from RPG and vice versa, I think we'll see that second group grow as well.<div>&nbsp;</div> PHP has more problems than just its i implementation. The implementation you have to take up with Zend, but no matter how good they do, the fact is that PHP is a scripting language that had to be fundamentally rewritten to graft OO onto it. It still doesn't support Unicode!<div>&nbsp;</div> The idea of thousands of PHP programmers is somewhat misleading. There may be thousands of people who have touched PHP, but at what level and for what purpose? A lot of the code out there is PHP4-based, and thus so are a lot of the programmers. PHP didn't get real OO capabilities until PHP5; PHP4 programmers won't know about that. And how many of those PHP programmers has any idea what a bill of materials or a three-way match is? The point being that as long as you limit your PHP to strictly UI coding, then PHP is probably an acceptable alternative. Not a good one, but an acceptable one.<div>&nbsp;</div> Now, PHP is free certainly. To program EGL, you need to get RDi-SOA, which is roughly $2000 a seat. Not cheap but it also provides a WYSIWYG JSF editor, a complete testing environment along with breakpoints, and of course complete integration with the i, including the ability to write RPG code for the back end. And remember, of that $2000, $800 is RDi (the RPG tooling), so the actual cost of EGL to an i shop is about $1200 a seat. Given how fast you can create an application in EGL, it's worth the price. Remember, tools aren't free: your time is. If EGL saves you 10 hours a week (and that's a very conservative amount), then is 500 hours of your time a year worth $1200?<div>&nbsp;</div> Finally, as to the web server, I assume you're talking about the memory on your i, since memory on a workstation isn't an issue these days. If you can't afford an extra GB of memory on your i, then you're going to have problems long-term anyway trying to move to the web. But if you do have 1GB available and you put that in its own dedicated pool for use by WebSphere, then I doubt that you'll see any performance problems provided you use EGL for the UI and RPG for the business logic.<div>&nbsp;</div> I hope that helps address some of your concerns, Chris. Personally, I'd rather IBM didn't spend another nickle on yet another technology to add to the mix and instead concentrated on getting the next release of EGL up and running. At the same time, before you make any decisions about which tool to use, I suggest you take a look at EGL. Download the trial, take a distance learning course, do whatever it takes to get a look at the package. After you've put together an AJAX page with RPG back end in 30 minutes, or added a Google map to a working Rich UI page with a single line of code, you really haven't seen the power of the language.<div>&nbsp;</div> Joe

3 shield commented Permalink

Joe<div>&nbsp;</div> I understand your commitment to EGL and if what you say is true I can see that it does have a lot of very good benefits especially its as easy to use as you say it is. OS/2 was far better and much more stable than Windows but guess what! BetaMax was a far better technology than VHS but guess what!<div>&nbsp;</div> Does EGL run on all platforms? PHP does, yes its a scripting language but probably runs most websites out there? Cant say the same about EGL yet... <div>&nbsp;</div> EGL sounds great, just as the i is a great server but we know how badly it is percieved in the market place now. Unless IBM puts it out there for people to use without a high price tag it will always struggle to be accepted. I don't program in RPG so the fact that it comes with RDB for RPG is of no benefit. I program on the i in C (I did see in the depths of the documentation that EGL can call ILE C so its not all lost). However I have paid for my WSDc licenses to be able to develop on the i, why should I now go out and buy more seats just to add EGL to it? I did see some information about upgrading using the WSDc licenses I have, but I have unlimited seats on the copy of WSDc I have today and I believe I have to go for a fixed seat option for this?<div>&nbsp;</div> If EGL is going to make my life easier I am all for picking up the books and getting going, however as far as I can tell its not widely accepted yet and I could spend an inordinate amount of time picking up skills which may never provide a return on that investment? <div>&nbsp;</div> I have to say PHP implementation on the i sucks, I installed Sugar CRM in my system as a test and found it took 12 - 15 seconds to provide responses where as my little Linux box took 1 - 2 seconds, the Linux box cost me about $600 and has 1TB of Raid5 DASD attached with 2GB memory. My i5 cost me $12,000 has 140GB DASD (No raid) and 1GB memory. Oh and I have never lost the Linux box and it runs 24x7 the same as the i5 (no I have not lost that either) BUT I do have to reboot the i5 more than the Linux box due to IBM software errors which is something I never did before. The i is now getting features and capabilities which make it more open and capable of working with the web etc, so this brings complexity and features which were not there before (Sorry to knock the i but it is not as stable these days).<div>&nbsp;</div> Now before you reply, I LOVE the i5, I don't develop on the Linux box anymore but could do quite easily. Linux does however provide me with the tools I need at a reasonable cost, if you had EGL for Linux I might be able to pick up the tool set (for free), try it out and learn it for investment of time only. Then move those skills to the i5 for the future. Unless IBM starts to give EGL away for free to get the momentum going and then provide more complex tools for a fee, I am not sure how well it will do, thats a pity because if its as good as you say, we should be looking at it. 30 day trials for development purposes dont work, I could not get to know any language in that short time especially if I am just looking at the capabilities for interest..<div>&nbsp;</div> I do applaud your commitment, I sincerely hope it takes off as well as you feel it should. I just feel the cost is out of my reach at this time.<div>&nbsp;</div> Chris...

4 JoePluta commented Permalink

Sorry, Chris, but I just don't understand your comments. You make statements that make it clear you haven't taken the time to even learn the slightest thing about EGL. For example, EGL runs everywhere. It generates Java code, so it's at least as portable as PHP, if not moreso - especially since PHP is very version-specific, and not every web server runs the same version.<div>&nbsp;</div> As to whether EGL will be around, I hope one of the IBM folks here will post to that effect. But again, with PHP you have to understand that there have been at least three different incompatible releases and from the looks of it, PHP6 will not be backwards compatible either, so that's four. So while PHP may well be around, your PHP4 code is all toast, as will be your PHP5 code when PHP6 comes out.<div>&nbsp;</div> You said something about paying for WDSC licenses, and then later you talked about unlimited options. Those two don't go together. You either have the standard edition, which came with the i and had unlimited seats, or you bought seats for WDSC/AE, in which case they could be transferred to seats of RDi-SOA (which includes EGL).<div>&nbsp;</div> EGL makes application programming easier, although as a C programmer, you may not appreciate it as much. EGL is a business language, not a system language, and as such might not have as much appeal to you as it does to RPG or COBOL programmers.<div>&nbsp;</div> Other things in your statement are strange, such as the non-RAID comment (why would you get an i without RAID?) and the lack of stability of the i. I don't want to turn this into an pro-i, anti-i rant, but there is still no system out there as reliable as the i. You want to run Linux, go ahead, but I'm not even a tiny bit interested. Linux is for appliances, not for business application development, and since this blog is devoted to EGL and the i (and specifically using EGL with RPG) I can't think you'd expect a lot of traction from arguments about the lack of stability or power in the i, especially in comparison to a $600 server running Linux.<div>&nbsp;</div> Any comparison of a dedicated Linux box to a multi-workload application server like the i is flawed from the beginning, because Linux doesn't have the built-in security, database, auditing, object-basing, single-level store, system functions, spooling, debugging and on and on and on...<div>&nbsp;</div> As to the trials or previews, I agree they should be longer. But then again, with the free distance learning that has been available, it's my absolute belief that 30 days with EGL will determine whether it's a fit for you or not. If you're a C programmer at heart who likes spending time working with pointers and allocating memory and converting between character codes, then EGL won't do a thing for you. If, on the other hand, you're a business application developer who just wants to create web pages and rich clients, easily pass data between your user interface and your business logic, write that business logic on the most powerful, stable and open application server in the world, provide code in multiple languages, have instant access to advanced Web 2.0 features and not have to learn Java or HTML or indeed any low-level language - well, EGL is for you.<div>&nbsp;</div> In short, if you are more interested in solving business problems than technology problems, then EGL will help you be the best you can be.<div>&nbsp;</div> Joe

5 shield commented Permalink

Joe<div>&nbsp;</div> Sorry it looks like I am ignorant, probably because I am, I did say right at the beginning I am an EGL newbie. The information I have been able to access has been mainly in your blog and I am just starting to unerstand where this all fits.<div>&nbsp;</div> Being a system programmer should not exclude EGL should it? I have system tools and utilities which could benefit from a web interface which PHP doesnt seem to work well for. Would EGL be a better fit?<div>&nbsp;</div> I dont expect any arguments about Linux and i, they are totally different, but the argument I get when questioning the cost is always about how stable it is, this is why we pay so much!<div>&nbsp;</div> I will keep kicking the tires until I understand EGL and how it can benefit me and my products. Changing from where I am today to a dedicated EGL programmer looking after application modernization should not be beyond my current skill set. If the industry is changing as we have been led to believe for a long time, having extra skills is not a burden. The problem is identifying the right skills!<div>&nbsp;</div> Chris...

6 JoePluta commented Permalink

You're not at all ignorant, Chris. It's just that you seem to be making arguments against EGL even though you really haven't had a chance to learn about it yet. That makes it a little harder for me to address the issues, because I don't know where it is that you just need a little more information, and where you have a legitimate question. But hey, I'm glad you're reading my blog! <grin>

Anyway, on to the question of system programming. While I'm sure the EGL folks expect that you could indeed write any type of application in EGL, you have to remember my focus: EGL and i, and specifically EGL and RPG or COBOL. I live and die in the business application spectrum, not the systems side. A lot of that is because if I really need systems work, I have libraries to use - either APIs on the i or the substantial array of Java libraries available for other tasks. If I need, for instance, to send an email, I just plug in a call to JavaMail and I'm done. So I don't really have to do a lot of system programming.
And of course, the very nature of EGL makes system programming unnecessary. The one area where I did do a lot of system programming over the years was in communications, character-set encoding, messaging, session management and all of the other complexities of a true multi-tiered architecture. With EGL, the vast majority of that coding is taken care of. That's what has drawn me so closely to the language - the 4GL concept of the record, with metadata, that can be passed from tier to tier and which acts as needed at each tier.
To put it succinctly, I don't know if EGL would be my first choice for system utilities such as the ones you write. At the same time, it would provide a powerful web interface and as I think about it, the Rich UI stuff in particular might be really nice for systems management consoles.
And of course if your clients are looking for application modernization and the ability to develop absolutely cutting edge Web 2.0 applications reusing existing business logic and without having to learn Java, HTML, and JavaScript frameworks, then EGL is simply the best tool to use.