Hacker News Reader

Comment by alkonaut | Hacker News Reader

Is there a modern and/or open alternative to this? E.g a SQLite + electron or local web client thing where you could build a simple inventory or similar but you should also be able to scale it to client server when the need occurs 10 years down.

Note that any number of cloud startups don’t count as an alternative to access. When these things start it’s as an excel sheet with data that no one will go through the enterprise hassle to get cleared for anywhere but their own hard drive. This needs to sit above excel but still local. You need the entire business case and business value while it’s still in a local directory.

View on HN

Arguably Saleforce. One aspect of the product is point and click editable database tables, reports, forms.

Actually I think the lesser known parts of office 365 'PowerApps' have various features for point and clicking your way to a cloud database, forms and reports, and even ETL from your existing data sources.

However, people don't think of these options like they think of access. For MS and Office 365, they could put more effort into marketing the features. Maybe they don't want to. For Salesforce, I guess they could try to get their product into non-computer people IT class syllabuses?

There's a huge push in MSFT for PowerApps.

Like ... crazy push. They're running free App in a Day workshops in hundreds of cities, and the "Power Platform" (Power BI, PowerApps, and Flow) is their key foot in the door sale in FY20.

Low code/no code is always a Modern Workplace linchpin.

NSBase[0] markets itself as Access replacement using sqlite as backend. It's built using Lazarus IDE and has lua scripting. It's free and works on Windows, Linux and Mac. I haven't taken it through the ropes to see how good it is.

[0]http://www.nsbase.neuts.fr/en/

Seems like it's Windows and Linux only, at least according to the (outdated) web, and the download page.

Thanks for correction. I cannot edit my comment, too much time has lapsed. I was reading up on Lazarus IDE about the time I discovered NSBase and somehow got it in my head that NSBase supports same platforms as Lazarus IDE. Apologies to those I have misinformed.

As soon as you bring HTML and JS in, projects choke and die. There is just too much goldbricking and fiddling and fussing and yak shaving around possible, and almost none of it is in service of actually making anything. The friction binds the gears up horrifically.

Something like WinForms with SQLite that didn't use Entity Framework would be far better. But really all we want is to get back to VB6 on top of Access. Maybe a few less footguns...

> As soon as you bring HTML and JS in, projects choke and die.

True! I wonder whether making JS event handler in little textarea boxes would actually help. Select “Even handler for field ‘Name’ on ‘click’”, then program your event. Obviously the end result is a js file, but they user is guided...^W prevented from doing something else than modifying output before submission.

Yep, building something like this at github.com/Zubairq/pilot but it needs docker to run

I have a good amount of experience with both winforms and modern web development. I would pick winforms over vanilla html and js. However, I would easily pick vuejs (or any comparable component based web framework) over winforms. Having maintained legacy VB apps with multiple 5000+ line forms, inscrutable multidirectional event chaining, and tangled up business logic and view logic, I've become a huge fan of forced separation of concerns.

The big problem is that there is much more of a learning curve. The first useful software I wrote used VB6 and winforms. Getting started was easy and intuitive, with a gentle learning curve. With virtually no experience, I was making neat things in hours. Within days I was making full applications. That says something about the power of that platform.

To be truly effective with modern web development took me a year or so. That was after years of professional experience.

At this point I can build web applications faster than I can build desktop software, but it took a huge time investment to reach this point, and I had to build multiple tools to make it possible.

I agree that there's a niche that needs filling. I want all the benefits of modern web development, and I want a learning curve like VB6 with Access.

> but you should also be able to scale it to client server when the need occurs 10 years down.

YAGNI. Why not just go with a process that you can't scale to that hypothetical use case, and once that process reaches maturity (stops being adjusted every week) and grows to the point scaling becomes a real consideration, and not just theoretical one - only then hire resources to migrate it to a new system that's scalable?

I think that Open/Libre-Office have options for this. You can also use Access as a front end for another db (odbc) backend pretty easily.

I'd probably do a web app myself, but that's just me. I know a lot of people that cut their programming teeth on Access apps, including distributed ones.

OO/LO "Base" is a pile crap in my opinion.

People often develop smallish CRUD apps in MS-Access in two weeks or less. It's up and going without fuss and muss and without fiddling with servers, containers, DBA's, etc.

You have to admire its nimbleness.

Yes, the database "crashes" fairly often, but it's easy to make frequent round-robin-style file-based backups using Windows Scheduler and DOS scripts.

Most databases cannot be properly backed up with file-based techniques because of syncing of internal pointers. But MS-Access's separate "lock-file" based technique somehow facilitates file-based backups.

Such MS-Access apps are far from perfect, but if you factor in everything, they seem to be a net benefit.

I'd probably do a web app myself

If follow-on maintainers don't know the web framework used, it can be hard to maintain. Web apps are rarely nimble to do without an involved framework on which the developer is familiar with. MS-Access avoids that problem with ubiquity and a light drag-and-drop learning curve. (It did get harder to learn when they went from coordinate-based to an HTML-esque flow engine around 2006.)

Let's face it, desktop GUI IDE's are usually much easier to learn than web stuff. You don't have to deal with the web's lack of state, and CSS/DOM/JS headaches/bugs/inconsistencies.

Whats wrong with Base? I haven't used it so I don't have any preconceptions about it.

Re: Whats wrong with Base?

When I first learned MS-Access, I barely had to read the manual. I could discover most functionality by clicking around. With Base, I had to constantly dig in the manual and still couldn't get it to do many ordinary CRUD idioms.

It's a clunky front-end with a fragile back-end. While it superficially resembles Access if you squint your eyes, it can't provide more than a small fraction of its utility currently. Too many missing/broken bits.

I spent a couple decades as a power user of the MS Office suite and find that I can generally make Calc do what I need as a spreadsheet and even Writer as a word processor. However, pretty much every time I've tried to develop a solution in Base it turns into a fight and I end up throwing up my hands and going with a different database solution. Calc doesn't even work all that well with it... and it's one of the few applications that works with it at all.

Yeah it doesn’t feel like something more than an access clone. I’d like something that is as easy as access but can also be migrated to a modern web+db app with ease. Even if it requires some minimal coding ability it would be cool to be able to make an Elm crud app deployable as a container, starting from an access-like set of tables and forms.

LibreOffice has Base which can use either embedded database system or talk to proper RDBMS like postgresql. KDE has Kexi which afaik is pretty similar. I have no experience with either, nor have I ever heard anyone to use them.

I'd probably do a web app myself

Running on? Wait, you mean I have to get IT to set up a server for me? Who's responsible for maintaining that? Backups? They're going to do that for something used by only 10 people that saves us maybe 20-30 hours a week? How much are they going to charge my department for this?

A company like Atlassian famously often had their first install at customers performed by employees on a rogue box under-the-table. Seems like it’s how one introduced software in the 2000s.

I wonder whether today’s equivalent could be made on Raspberry Pi. It requires a few tricks, but there is so much demand from managers who could afford paying 50€ out of pocket.

Of course, having a rogue box plugged to the internal network with a Wifi and Bluetooth chip on board is not IT-approved. Can it be bypassed?

Check out Oracle application Express (APEX). HN loves to hate Oracle, but it is a great tool for a building crud applications and dashboards.

This. Its interesting to read this thread mentioning DB and systems issues that are solves in Oracle APEX. We built an ERP system for a oil&gas client, including in the requirements, apps for monitoring gas values at the plant, by simply connecting to the mesurment equipment API. All that on Oracle XE, free version that comes with APEX. Three weeks to build apps, five months to implement in the company. I'm thinking software is not the problem, its people, politics in the company, resistance to change, really everthing but tech. We're solving business problems, helping the company cut costs, increase speed and streamline processes. Tech is a tool - there are so many other obstacles to solving problems other then tech. Thanks for mentioning Apex, and Oracle, in a thread about DB-s that fails to mention the company building databases for 40 years, that run for instance at CERN, to name but one.

At this point, I think APEX is the best product and biggest selling point of Oracle. Too bad they don't focus in selling it.

The largest drawback is that it comes with an Oracle database bundled, so many of the savings in development time are wasted dealing with it.

Not that I've seen. You figure with file base databases it wouldn't be hard to put something together (sqlite?)

Libre office has "base" but its just a front end. My boss uses access, she like the query builder. I've got my mysql instance and some front end tools that do the same, but I'm a developer.

Filemaker is another of these applications, but not open source. The thing these stand alone apps have is direct printer access for printing forms and labels. It makes it hard to get off.

(Our get off filemaker solution is to download the data from our website and use a stand alone print app to print the labels. One extra step, but its not completely ideal.)

I would love to build one some day. I have ventured into building a complete personal finance dashboard that was purely local, but lost steam in the investigation phase. I am guessing there is a market for desktop applications that can do these things once privacy becomes more important.

VBA in the current version of office. Keep the databases in MariaDB. Scary but it works.

Maybe Filemaker / Claris

https://www.filemaker.com

I think Python could be a good alternative. It has modules in stdlib to interface with SQLite and Tk if you want to do an (ugly) user interface quickly. It does lack a good user interface to manipulate the database, though.

Airtable, maybe?

Reply on news.ycombinator.com