Nothing important, except...

My ERP5 story

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License: http://creativecommons.org/licenses/by-nc-nd/3.0/

A bit of background -- "in the beginning the earth was without a form and a void". Not earth, but one company in Poland (you can find its name in archives of erp5-users and erp5-dev ;) ) needed ERP system, but was not able to find any. I had an opportunity to be "the chosen one" to implement such system. I get hired and we stated "let us set up full ERP configuration in 6 months". Wow, 6 months, it will take you soooooo long? -- someone can ask. Indeed, this is huge amount of time comparing to implementing LDAP or SAMBA architecture, but those are pure technical systems (even if complex, this is technical, engineering complexity) and ERP is "business" system. So we assumed that 6 months is enough. Ah, indeed, what we want to achieve during that time? Flexible MRP/Trade solution, with all its aspects (customers, suppliers, accounting, invoicing, BOMs, supply chains, production, consumption, forecasting...).

We were wrong about implementation time, fortunately we assumed some margin, well read on.

So I started to work around implementing ERP solution. Let us skip the part of choosing the most suitable ERP open source based solution (ERP5 won after dramatic battle!) and jump to first implementation.

What I did? I took the simplest working implementation that times -- download and burn LiveCD and start to play. After becoming a bit familiar with Zope, python and ERP5 I again followed "the easy way". I checked which Linux distribution is supported by ERP5 -- it was Mandriva -- and I switched my laptop to use it. Thanks to this, after some minor issues, I was able to "run" ERP5 -- install Business Templates, hack and play, and after each SiteError I couldn't recover -- remove and reinstall...

...it was few years ago, and what today if I would be a total newbie without any experience in Zope, python and ERP5? I wouldn't use buildout for first try, I would use TioLive -- in online form or as a grid.tiolive.com. Even if not perfect, it would solve my issue -- to see and LEARN ERP5 which is running without taking care about low level technical aspects as BUSINESS aspects are more important. That time we understood that we need as integrated as possible ERP system, not yet-another-set-of-technical-tools-which-put-some-data-in-relations. I would choose the shortest possible path to start LEARNING what ERP and ERP5 are. Today there is much more documentation about it then few years ago. I feel sad, that people are not going on such path, but well, everybody is free to choose whatever she/he wants to do.

So story continues, after few months of "step forward, two steps backward, three steps left, ..." we had running site more or less configured. Ah, indeed, we hadn't read about what is the 5 in ERP5, but something was working. Hopefully we found possible support in Poland - mr. Bartek Gorny was providing it -- as far as I recall correctly he was home-grown ERP5 hacker that time. He was really happy with using Zope and ERP5 and we were able to buy his time to provide us support. Wow, it was good, to see someone else working in environment we tried to create. His training boosted us and helped to play with many technical and business issues in ERP5.

Later Bartek arranged a meeting with Jean Paul in Poland. We told him about our assumptions (we need it in 6 months, ble, ble) we've shown what we have and he replied:

And well...he was right. Again, we assumed that ERP5 is technical tool and we have technical problem to solve. No way, we had business to run and ERP5 is business tool. So it requires business knowledge (it requires technical knowledge too, but it is nothing, as it ends up with working and not configured for business ERP5 site...).

So we reviewed our goals and set them from pure business point of view, separate implementation steps and extended implementation time. Of course we dropped current achievements, as they were not fitting at all with our reality.

Again few months passed, and we did what here, on this list is so criticised. I was sent to Nexedi office, my boss paid for this. Just one week, but wow, I tell you, wow. It speed up our project for about 6 months. People here were really cool, everybody tried to explain me everything with great patience. Even if my questions were "stupid" (and most of them were like this, today I know). After one week of hard work (and a lot of beers, guys, beers here are great!) I went back and worked like a crazy guy for another amount of time (again in months). After about one year of my journey with ERP5 I started to contribute something -- small pieces of code, mostly tests, etc. Time passed, company which hired me was (and still is ;) ) happy, system become more and more suited. It still runs and it is still used. It still runs, because it is tested -- I agree with the fact that "what is not tested IS broken for sure, works only by chance". I love tests, they are cool, they run automatically. And they are computer's language representation of bug reports. And we are geeks, aren't we? ;) So let us use commonly known language to describe issues and profit from fact of running tests automatically again and again.

So indeed, it is POSSIBLE to implement ERP5 as "external" not related developer. It requires some effort and work, but from my experience -- patience and realistic goals are critical.

Sweet, isn't it? Yep, but not perfect. I did a lot of mistakes, which I wouldn't do today. For example I did too advanced things too early, which ended up with not-so-perfect implementation (but still working). I was not able to use preconfigured thing like TioLive so our system is not so clean as it would be (that is why I am stressing this so much, start with TioLive and stay there for some time, play, run your or fictional company). I lost most of my time because of focusing too much on technical instead of business aspects and quite often wrongly designed idea was implemented perfectly (from technical point of view), than have to be dropped, to have proper design but with poor implementation, because time was already consumed.

So that is it. I am example of the fact that it is possible to implement ERP5 alone, without a big community.

ŁukaszNowak (last edited 2011-01-17 16:54:44 by ŁukaszNowak)