If you support z/OS and have access to the Internet, you almost certainly have been impacted by a sweeping website redesign that IBM is currently implementing. To say that we, and much of the z/OS community, are unhappy is a massive under-statement – if you don’t normally follow IBM-MAIN, you should take a few minutes to see what customers are saying about how these changes have affected them. Sadly, we have no control over what IBM does. And based on the feedback from our customers, it seems that even the community of customers that collectively pay IBM billions of dollars for IBM Z hardware, software, and related services were not consulted prior to these changes being rolled out. We can’t change what has already happened, but maybe we can help a little to reduce some of the disruption that these changes has caused.
What Exactly Is Happening?
The powers-that-be in IBM decided to redesign and update the www.ibm.com website. One of the reasons is to reduce the number of outdated, unused, and duplicate pages from the hundred million URLs that IBM manages. The intention was to make searches more meaningful. We applaud this initiative and can certainly understand why pages on MVS/XA and OS/390 only muddy the (search) waters.
There is also talk about IBM (or some parts of IBM) wanting to use the ‘ibm.com domain’ primarily for marketing purposes, so that objective is probably behind some of the changes that we are seeing as well.
The intent was that these changes should be largely transparent to the users of the IBM sites, with redirects being set up to protect users from broken links. However, it appears that implementing these changes resulted in some technical material being removed or moved to other locations, with missing or nonsensical redirects.
Piecing together information received from our customers, it appears that some of the IBM teams on the ground simply didn’t have enough time to create link redirects for the pages they owned. As a result, links from other IBM sites, other non-IBM sites (including ours), Redbooks, our Tuning Letters, and countless other sources no longer work. Google does not appear to be much help, either because it can’t find the new home for this material, or because it is simply gone – we can’t tell.
The pain caused by the broken links has been compounded by the fact that there doesn’t appear to be a consistent strategy about where material should be moved to. As a result, it appears to have been left to the owners of those sites to try to find some other home, either on some other IBM domain, or on the Microsoft-owned GitHub. In the performance area (WLM, RMF, SMF, pricing), I’ve found that two-thirds of my bookmarks no longer work.
In addition to this mass deletion and/or re-housing, there are conflicting rumors about an initiative to delete infrequently-referenced technical material over a certain age. As part of a cleanup of obsolete material, it is understandable that some older (but potentially still valuable) material was deleted (including some of Frank’s early Redbooks!). But just because material is old and maybe even infrequently referenced does not mean that it is obsolete. We hope that the people (or robots) doing this ‘cleanup’ factor in the size of the mainframe population compared to the PC population when deciding if something is infrequently referenced – ‘infrequent’ is a very relative term. In any case, we have been told by IBM that if readers can identify valuable information that was deleted, they will restore it. Of course, the challenge is that you can’t ask for something if you are not aware that it existed in the first place.
Also, as part of this cleanup, the Knowledge Center pages that contained all the z/OS 2.1 material disappeared last week. This is even though z/OS 2.1 is still in support and most of the Google links that point to z/OS manuals currently point at the 2.1 manuals. As a result, just about all the search engine links to the z/OS product manuals were broken. To make matters worse, the “Change version” button that was supposed to bring you to the corresponding page in the other selected release was also broken.
Part of the response to the changes on the ibm.com domain appears to have been to move some of the z technical materials from their previous homes to the IBM ResourceLink site or the IBM developerWorks site. This allows that material to be retained; however, during the transition, we are finding many broken links and links to the wrong pages. Also, some of the new ResourceLink pages require the user to have to sign on to ResourceLink (this is not intended and IBM are working to resolve this). And the developerWorks site requires that you understand communities. You also appear to need an IBM ID in order to access developerWorks (I just tried, and I was redirected to the signon page). “z/OS” isn’t even mentioned on the home page.
The disappearance of the KC z/OS 2.1 material highlighted the fact that there ARE alternatives. One option is to install Knowledge Center for z on your z/OS system, so you have your own copy of the manuals, under your control. Also, the z/OS 2.1 material is still available in PDF format – you can find it here:
https://www-304.ibm.com/servers/resourcelink/svc00100.nsf/pages/zosv2r1-pdf-download?OpenDocument. However, the problem that affected most people was that the search engines point at the KC version of the manuals, and all the z/OS 2.1 manuals were removed from KC. So having KC for z, or access to the PDFs means that you can access the information (if you know which manual to look in), but it doesn’t help with the search engine problem.
We’ve talked to the z/OS KC team and they have now restored the 2.1 material into KC. The problem with the Change version function has also now been fixed. They also plan to redirect the Google links to the z/OS 2.3 KC before considering removing them again.
What Is The Point Of All This Disruption?
Like 99.9% of the world’s population, I use a search engine to find the information I need to do my job. If it is a website that I visit frequently, I will add it to my bookmarks. Whether the URL is www.ibm.com or developer.ibm.com or www.research.ibm.com or whatever, is utterly irrelevant to me. I simply want to find the information I need to do my job. So I don’t understand what IBM expects to achieve by moving material from one IBM site to another? In the age of search engines, trying to assemble all the IBM marketing material on one site seems like a labor-intensive, disruptive, and pointless exercise that will not ultimately benefit either IBM or its customers.
For the last 3 years, it seems that no IBM executive can make a presentation without talking about ‘Agile Development’ and ‘Design Thinking’, and about how these are really great because now IBM speaks to customers before they go and create something. I think they may have missed this step in the website redesign. In the words of one of the contributors to IBM-MAIN, “This is sabotage! Sabotage to IBM’s clients!”
It has been left to two very patient IBMers, John Eells and Sue Shumway, to take the customer slings and arrows and offer their personal apologies and potential work-arounds on IBM-MAIN. But the long delay in providing any sort of official statement from IBM indicating that they were aware of the impact of their actions, or their plans to undo the damage, just made people even more frustrated. We hope that IBM will correct this and learn from the experience.
How Will This Affect You?
There are multiple impacts. The most lasting one, probably, is this decision to delete material that someone has decided is no longer of value. Hopefully IBM has only removed the documents in question from its website, and not actually deleted the material itself. So, if you are looking for something that you know USED to be there, speak to your account team and maybe they can find some kind soul in IBM that will help. But for information that used to be out there that you were not previously aware of, that is effectively lost to you forever – how can you ask for a copy of something that you don’t know exists?
This might come as a shock to some people that are driving this initiative, but not all IBM products were created within the last 5-10 years – especially the ones that generate the bulk of the z/OS-related software revenue. Most of the really useful information about a product is probably created within the first few years after a product is introduced. So, if you are looking for good technical information about z/OS, CICS, DB2, IMS, MQ, JES2, RACF, ISPF, WLM, or WAS, your chances of success have just decreased.
IBM says that this project has been progressing about 18 months with pages still being changed. There is still lots of useful technical information on the IBM site, but we have no idea if that is because a conscious decision was made to leave it there, or if the cull just hasn’t reached those pages yet. The best advice we can offer is to download any useful information you find as soon as you find it.
Thankfully, the issue with the search engines and the removal of the z/OS 2.1 material from KC has now been addressed. And hopefully the force of customer reaction will ensure that the eventual removal of the z/OS 2.1 material will go much smoother.
Other issues? Bookmarks. We all have long lists of bookmarks of our favorite sites. Links that were broken by the removal of the z/OS 2.1 KC material should be working again now. However, bookmarks for many invaluable technical pages are broken and will likely remain so until you take some manual action, or until IBM provides redirects to the correct new location. We do know that IBM is actively working to identify pages that have been moved, and plans to provide redirects where appropriate.
Having said that, I don’t think that we can expect that every single link that you might have bookmarked will be ‘fixed’. Therefore, to help our subscribers that find themselves in this situation, we have set up a web page that contains a list of the new URLs. It is far from comprehensive – there are simply too many web pages to cover. But we hope that we have captured the most popular ones. If you have a URL that has disappeared, and you can’t find its new home, email us and we will try to help you find it.
Closing Note
We are continuing to work on this. We have been speaking to IBM, and the z people are very aware of customer feeling on this topic. They are addressing the items that are within their control and have asked us to help them identify outstanding broken links and missing information. We hope that people further up the ladder in IBM will grasp how outraged customers are over the decision to make ibm.com mainly marketing – and how that decision will adversely impact IBM itself. We strongly encourage all our readers to let your IT executives and your IBM account team know your opinion on this change – your voices are the best hope we have to at least limit further damage.
They say that ‘it is an ill wind that brings no good’. We do know that this event has greatly increased IBM’s awareness of how heavily customers rely on search engines to access material on IBM websites, so hopefully that will positively influence their actions in the future. This focus on KC has also resulted in getting the Change version bug fixed. We hope that the KC 2.1 experience will encourage IBM to find some way to access publications that is release-independent in future. We believe that all this focus will also provide an opportunity to update existing pages that had genuinely out-of-date information (such as the zFavorites page). And this situation also gave me an opportunity to try out the “Adobe Indexed PDF collection” – it is really cool, you should give it a try if you haven’t already.
But there is one very important final point that we want to make. We understand how frustrated people are, and how these changes are having a huge adverse effect on your productivity. But please remember that the people that you are likely dealing with in IBM are not responsible for these impactful changes. They are feeling the same pain as you, and it is only heightened when customers vent their frustrations on them over things that they have zero control over. By all means, make your feelings known to your own executives and to your IBM account team. But please be kind to your IBMers – being a techie in today’s IBM is not much fun at the best of times, so show a little good humor and acknowledge that we are all in the same boat together.
We will continue to work with IBM and with our customers to reduce the impact of these changes. If there are valuable pages or information that you can no longer find, please let us know and we will pass the information along to the IBM Z team. Similarly, if we get updates from IBM on their plans to restore critical missing information, we will pass that information along to you.