What does Legacy really mean?

Posted on Updated on

In the IT industry we hear the term “legacy” being using, but that does it mean? It can mean a lot of different things and it really depends on the person who is saying it, their context, what they want to portray and their intended meaning. In a lot of cases people seem to use it without knowing the meaning or the impact it can have. This can result in negative impact and not in the way the person intended.

Before looking at some (and there can be lots) possible meanings, lets have a look at what one person said recently.

“Migrating away from legacy databases like Oracle can seem like a daunting undertaking for businesses. But it doesn’t have to be.”

To give context to this quote, the person works for a company selling products, services, support, etc for PostgreSQL and wants everyone to move to ProtgreSQL (Postgres), which is understandable given their role. There’s nothing wrong with trying to convince people/companies to use software that you sell lots of services and additional software to support it. What is interesting is they used the work “legacy”.

Legacy can mean lots of different things to different people. Here are some examples of how legacy is used within the IT industry.

  • The product is old and out of date
  • The product has no relevancy in software industry today
  • Software or hardware that has been superseded
  • Any software that has just been released (yes I’ve come across this use)
  • Outdated computing software and/or hardware that is still in use. The system still meets the needs it was originally designed for, but doesn’t allow for growth
  • Anything in production
  • Software that has come to an end of life with no updates, patching and/or no product roadmap

Going back to the quote given above, let’s look a little closer at their intended use. As we can see from the list above the use of the word “legacy” can be used in derogatory way and can try to make one software appear better then it’s old, out of date, not current, hard to use, etc competitor.

If you were to do a side-by-side comparison of PostgreSQL and Oracle, there would be a lot of the same or very similar features. But there are differences too and this, in PostgreSQL case, we see various vendors offering add-on software you can pay for. This is kind of similar with Oracle where you need to license various add-ons, or if you are using a Cloud offering it may come as part of the package. On a features comparison level when these are similar, saying one is “legacy” doesn’t seem right. Maybe its about how old the software is, as in legacy being old software. The first release of Oracle was 1979 and we now get yearly update releases (previously it could be every 2-4 years). PostgresSQL, or its previous names date back to 1974 with the first release of Ingres, which later evolved to Postgres in early 1980s, and took on the new name of PostgreSQL in 1996. Are both products today still the same as what they had in the 1970s, 1980s, 1990s, etc. The simple answer is No, they have both evolved and matured since then. Based on this can we say PostgreSQL is legacy or is more of a Legacy product than Oracle Database which was released in 1979 (5 years after Ingres)? Yes we can.

I’m still very confused by the quote (given above) as to what “legacy” might mean, in their scenario. Apart from and (trying) to ignore the derogatory aspect of “they” are old and out of date, and look at us we are new and better, it is difficult to see what they are trying to achieve.

In a similar example on a LinkedIn discussion where one person said MongoDB was legacy, was a little surprising. MongoDB is very good at what it does and has a small number of use cases. The problem with MongoDB is it is used in scenarios when it shouldn’t be used and just causes too many data architecture problems. For me, the main problem driving these issues is how software engineering and programming is taught in Universities (and other third level institutions). They are focused on JavaScript which makes using MongoDB so so easy. And its’ Agile, and the data model can constantly change. This is great, up until you need to use that data. Then it becomes a nightmare.

Getting back to saying MongoDB is legacy, again comes back to the person saying it. They work at a company who is selling cloud based data engineering and analytic services. Is using cloud services the only thing people should be using? For me it is No but a hybrid cloud and on-premises approach will work based for most. Some of the industry analysts are now promoting this, saying vendors offering both will succeed into the future, where does only offering cloud based services will have limited growth, unless the adapt now.

What about other types legacy software applications. Here is an example Stew Ashton posted on Twitter. “I once had a colleague who argued, in writing, that changing the dev stack had the advantage of forcing a rewrite of “legacy applications” – which he had coded the previous year! Either he thought he had greatly improved, or he wanted guaranteed job security”

There are lots and lots of more examples out there and perhaps you will encounter some when you are attending presentations or sales pitches from various vendors. If you hear, then saying one product is “legacy” get them to define their meaning of it and to give specific examples to illustrate it. Does their meaning match with one from the list given above, or something else. Are they just using the word to make another product appear inferior without knowing the meaning or the differences in the product? Their intended meaning within their context is what defines their meaning, which may be different to yours.